- Email: [email protected]
Carotid endarterectomy: A safe cost-efficient approach
Carotid endarterectomy: A safe cost-efficient approach Paul E. Collier, M D , Sewickley, Pa. The diagnosis-related groups have encouraged physicians to become more efficient in the care of their patients; often, however, raising the question of safety. For 3 years all patients undergoing carotid endarterectomy at our institution were monitored in the intensive care Unit for 24 hours and the majority were discharged on the second postoperative day. After review of these patient's hospital records and direct patient interviews, it was clear that many patients did not require a stay in the intensive care unit and could be discharged on the first postoperative day. In January 1991 a prospective policy was established to evaluate the safety and efficacyof outpatient arteriography, same-day admission, selective use of the intensive care unit, and early discharge on the first postoperative day when feasible. During a 10-month period all patients undergoing carotid endarterectomy at our institution were evaluated (n = 52). Eleven patients had had a prior stroke (21%), 31 had either amaurosis fugax or transient ischemic attacks (60%), and 10 had no symptoms (19%). The arteriogram for 49 of the patients was obtained on an outpatient basis or during a prior admission, and these patients were admitted to the hospital on the day of operation. Nine patients were placed under general anesthesia and had shunting procedures, and 43 patients had cervical block anesthesia, eight of whom had shunting (19%). Only five patients required an intensive care unit stay for either hypertension, hypotension, or neurologic complication (one transient ischemic attack and one minor stroke). Forty-six patients (88%) were discharged on the first postoperative day; average length of stay was 1.29 days/patient. All patients were seen 3 to 5 days after discharge. There were no readmissions for hypertension, hemorrhage, or cardiac or neurologic events. There was no 30-day mortality. The hospital charge was more than $1900 below the reimbursement for the diagnosis-related group on average. From this series it is concluded that with this protocol, a short stay with selective use of the intensive care unit is safe and cost-effective for patients undergoing carotid endarterectomy. (J VASC SURG 1992;16:926-33.)
The introduction of the diagnosis-related group (DRG) system challenged hospitals and physicians to scrutinize their efficiency in delivering services. One of the most dramatic results of prospective payment has been the gradual reduction in the hospital length of stay (LOS) for many surgical procedures. This has also led to a reduction in the cost of delivering these same services. Studies such as the North American Symptomatic Carotid Endarterectomy Trial and Veterans Administration cooperative studies have shown that carotid endarterectomy can be performed safely in experienced hands. 1'2 However, questions remain about From the Department of Surgery, SewickleyValley Hospital, Sewickley. Presented at the Sixth Annual Meeting of the Eastern Vascular Society,New York, N.Y., April 30-May3, 1992. Reprint requests: PaulE. Collier,MD, Suite200, 301 Ohio River Blvd., Sewickley,PA 15143. 24/6/40729 926
what an appropriate LOS for a patient after a carotid endarterectomy should be and whether there will be an increase in morbidity or mortality with earlier discharge. Originally patients tentatively admitted to the hospital for carotid endarterectomy at this institution were admitted the clay before operation for arteriography unless they were already in the hospital. If the arteriogram showed an operable lesion, the patient underwent a carotid endarterectomy the following day. The majority of patients were operated on with the use of cervical block anesthesia and selective shunting. Postoperatively, all patients were monitored in the intensive care unit (ICU) for 24 hours and then discharged on the second or third postoperative day. After reviewing the hospital records of all patients who underwent carotid endarterectomy between 1988 and 1990 and directly interviewing many of these patients, the following hypotheses were made.
Volume 16 Number 6 December 1992
Arteriography could safely be performed on an outpatient basis before carotid endarterectomy. Patients undergoing carotid endarterectomy could be admitted to the hospital on the day of operation and the use of the ICU postoperatively could be limited to select patients. Finally, it was hypothesized that patients could be discharged early, often on the first postoperative day, without any adverse effects. A protocol was established to prospectively evaluate these hypotheses. The results of that study are the basis of this report.
PATIENTS AND METHODS Between January 1 and October 31, 1991, 51 patients underwent 52 carotid endarterectomies at Sewickley Valley Hospital. There were 36 men and 15 women ranging in age from 44 to 84 years. All 51 patients were entered in the following protocol. Unless the patient was in the hospital, a history, physical examination, and review of the duplex scan were performed in the office. If the patient was thought to be a suitable candidate for carotid endarterectomy, an outpatient arteriogram was scheduled. During this initial patient encounter the technique and possible complications of arteriography were discussed with the patient. More important, the feasibility of outpatient arteriography and the safety record of the radiology department were stressed so that the patient would be comfortable with the outpatient procedure. If the arteriogram demonstrated an operable lesion, the patient was scheduled for operation when he or she was next seen in the office. At that time, the risks and possible complications of the operation were discussed with the patient, and whenever feasible, a family member was present during this discussion. During this encounter our results and the possibility of early discharge were discussed with the patient so he or she could make arrangements ahead of time and feel comfortable at the time of discharge. Whenever possible, cervical block was used for anesthesia and shunting was done selectively. Patients were closely monitored in the recovery room for a minimum of 4 hours. Blood pressure was monitored with indwelling arterial lines in all patients. If the patients were unchanged neurologically, had no cardiac problems, and maintained a systolic blood pressure less than 180 mm H g without medication, they were transferred to the vascular floor; otherwise, they were monitored in the ICU. Patients were discharged on the first postoperative day if their neurologic status was unchanged, there were no problems with their blood pressure or heart, their wound was satisfactory,
Carotid endarterectomy 927
and they felt comfortable going home at that time. All patients were seen in the office between 3 and 5 days postoperatively. Of the 51 patients undergoing carotid endarterectomy, 26 (51%) were receiving antihypertensive medication, seven (14%) were diabetic, and 23 (45%) were smokers. Twenty-seven (53%) had coronary artery disease manifested by either angina pectoris, a history of myocardial infarction, or a prior coronary artery bypass. Thirty-one (60%) of the operations were performed for transient ischemic attacks or amaurosis fugax, 11 (21%) for a prior stroke, and 10 (19%) for an asymptomatic stenosis greater than 80%. Forty-three carotid endarterectomies were performed with cervical block anesthesia. Eight (19%) of these patients required shunts. Nine patients required general anesthesia and all of these had shunts. The reasons for general anesthesia included recent stroke, patient anxiety, and one late reoperation for a recurrent, symptomatic stenosis. RESULTS Three arteriograms were performed on patients who were already in the hospital. Of the 48 other arteriograms all were successfully performed as outpatient procedures. The only complication encountered was a small groin hematoma, which resolved with conservative therapy. There were no neurologic, cardiac, or renal complications. None of the 48 patients required admission to the hospital as a result of the angiogram. Two neurologic events occurred as a result of the 52 carotid endarterectomies. One patient had weakness of the left arm that lasted less than 5 minutes while the patient was in the recovery room; there was subsequent total recovery. Duplex scanning demonstrated a patent repair with no defects. A second patient had undergone carotid endarterectomy after a stroke with near total resolution 6 weeks previously. After she had vomited in the recovery room, bradycardia and hypotension developed with recurrence of the prior stroke symptoms. Restoring her blood pressure did not change her symptoms so she was rapidly returned to the operating room. No thrombus or technical defect was found at reexploration. She was left with residual symptoms, worse than she had preoperatively. She now functions independently with minimal residual symptoms. For this group of 52 operations, the permanent stroke rate was 1.9%. Both neurologic events occurred within 2 hours after the operation. No neurologic events occurred after the patients were discharged from the recovery room. Transient problems with hypertension were corn-
Journal of VASCULAR SURGERY
928 Collier
mon. Twenty-eight patients (54%) required continuous intravenous nitroglycerin administration to maintain a systolic blood pressure less than 180 mm Hg intraoperatively. Twenty-six patients (50%) required antihypertensive treatment in the recovery room. Of the 24 patients whose blood pressure was controlled in the recovery room, 23 (96%) had therapy terminated within 2 hours. The other patient's treatment lasted 3 hours. Only two patients (4%) did not have their blood pressure normalized while in the recovery room and required further treatment and monitoring in the ICU. Antihypertensive therapy was required for 8 and 9 hours in these two patients. Persistent hypotension developed in two patients (systolic blood pressure less than 100 mm Hg). One was monitored in the ICU overnight and the other was monitored on a telemetry floor without any adverse effects. Retrospectively, neither patient required active ICU treatment. Only five of the 52 patients (10%) required postoperative monitoring in the ICU: two for neurologic changes, two for persistent hypertension, and one for persistent hypotension. No patient remained in the ICU for more than 24 hours. Forty-six patients (88%) were discharged after only 1 day in t h e hospital. Two patients were discharged on the second postoperative day, two on the fourth, and one each on the third and sixth postoperative days. The average LOS was 1.29 days per operation. The average hospital charge was $3408 per patient admission. DRG 005 (Extracranial Vascular Procedures) reimburses the hospital $5396. This resulted in an average profit to the hospital of $1988 for each carotid endarterectomy performed during the 10-month study period. There were no readmissions or emergency department visits for hypertension, hemorrhage, or cardiac or neurologic events. Because of a quirk in the Medicare rules and regulations, an outpatient arteriogram is reimbursed outside of the DRG if it is performed more than 72 hours before the hospital admission. These data were therefore retrospectively reevaluated to determine whether there was an overall cost reduction for the system following this protocol as compared with performing an inpatient arteriogram and extending the length of stay by 1 day. For Medicare patients this protocol cost the system $900 more per patient when compared with inpatient arteriography. From a social standpoint, however, this protocol provides improved bed utilization for sick patients in a hospital
that is often operating at full capacity. Because reimbursement by all other payers, like Blue Cross, is based on actual costs and charges, it was found that the protocol that used outpatient arteriography was slightly less expensive because the cost of the short stay unit was less than the cost of an extra day in the hospital. After discharge the patients were interviewed regarding their satisfaction with the protocol. None of the patients believed that they had been forced to leave the hospital before they were ready. Most patients thought they did better at home. Only three patients complained that they would have preferred to stay in the hospital after their arteriogram, but they were happy to be home after their operation. DISCUSSION
The results of this study demonstrate that the protocol used is safe for the patient and cost-efficient for the hospital. Some of this efficiency is secondary to the cooperation between the emergency physicians, primary care physicians, radiologists, and vascular surgeons all working in one institution. Patients with transient neurologic symptoms and normal computerized tomographic scans are not routinely admitted to the hospital. They undergo an expeditious outpatient workup and are then admitted if they are a candidate for operation. The shifting of the arteriogram from the inpatient to the outpatient setting accomplishes two ends. First, it removes the expense of the arteriogram from the hospital cost that is reimbursed by DRG 005, and second, it reduces the hospital LOS by 1 day. The cost analysis of carotid endarterectomy done by Maini et al.3 included the arteriogram as part of the inpatient admission. This lengthened the reported LOS and cost. In our study there were no complications in any of the selective carotid arteriograms performed. In our study digital subtraction intraarterial techniques were used to reduce the dye load, and direct carotid injections were only used if unsatisfactory pictures were obtained. Limiting the use of direct carotid injections reduces the risk of procedure-related stroke. Current technology allows the use of very small (4F or 5F) catheters for arterial cannulation, which therefore reduces the risk of hematoma. The skill of the radiologists has virtually eliminated local arterial complications. These results clearly demonstrate the safety of outpatient carotid arteriography in both patients with symptoms and those without symptoms.
Volume 16 Number 6 December 1992
Same-day hospital admission for carotid endarterectomy procedures is possible only if the patient's blood pressure is well controlled and their cardiac status is satisfactory. Corson et al.4 showed that perioperative hypertension correlated best with uncontrolled preoperative hypertension. They believed that good preoperative blood pressure control would provide better perioperative control. This is demonstrated by the current series in which all patients had well-regulated blood pressure before admission, with the result that only two patients required ICU treatment for prolonged hypertension. Cardiac status is evaluated by history and resting electrocardiogram. Dipyridamolethallium scanning is obtained only if the clinical picture or electrocardiogram warrant it. In light of the reported mortality rates of less than 1% for carotid endarterectomy with regional anesthesia,5,6 dipyridamole-thallium stress tests are reserved for patients with worrisome cardiac symptoms. The 0% myocardial infarction rate in this series again attests to the low cardiac risk for carotid endarterectomy done with regional block. Regional anesthesia was used in the majority of patients for three reasons. First, as shown by Corson et al., 4 when perioperative hypertension occurs, it is of shorter duration when regional anesthesia is used. In their series postoperative hypertension averaged 1.4 hours when regional anesthesia was used and 19.6 hours when general anesthesia was used. In the current series only two patients had hypertension lasting more than 3 hours postoperatively that therefore required ICU care. One patient had been operated on with general anesthesia and the other with regional block. The second reason for using regional anesthesia is the extremely low myocardial infarction and mortality rates reported? ,S,6 Finally, regional anesthesia is used because it provides the best method of neurologic monitoring during carotid cross-clamping. Selective use of the ICU after carotid endarterectomy has been shown to be reasonable and cost effective by O'Brien and Ricotta. 7 In their series only 18% of patients required a stay in the ICU for an average of 24.5 hours. Their study showed that those patients requiring ICU care can be identified in the early postoperative period. They showed that the charge for the ICU added 12.5% to the patient's bill. The two myocardial infarctions in their series occurred 2 hours and 6 hours postoperatively. In the current series there were no myocardial infarctions
Carotid endarterectomy
929
and both neurologic events occurred while the patients were still in the recovery room. The safety of this selective policy is demonstrated by the fact that there were no late admissions or readmissions to the ICU. Because only 10% of the patients required ICU stays in this series, 90% of the patient's hospital charges were lower by 12.5%. The LOS for most procedures including carotid endarterectomy has been reduced by the introduction of prospective payment? In this series the average LOS was reduced to 1.29 days per patient. This was done by the use of outpatient arteriography, same-day hospital admission for the procedure, regional anesthesia, and early discharge when feasible. The cost efficiency of this approach is clear. The safety is demonstrated by the fact that there were no returns to the hospital for uncontrolled hypertension, bleeding, or neurologic or cardiac problems, and there were no mortalities within 30 days of operation. Multiple factors contribute to the safety and efficiency of carotid endarterectomy. From the data presented in this study it is concluded that this protocol of regional anesthesia, selective use of the ICU, and early hospital discharge is a safe and extremely cost-effective method of treatment for patients undergoing carotid endarterectomy. REFERENCES 1. North American Symptomatic Carotid Endarterectomy Trial (NASCET): North American Symptomatic Carotid Endarterectomy Trial: methods, patient characteristics and progress. Stroke 1991;22:711-20. 2. Towne JB, Weiss DG, Hobson RW II. First phase report of cooperative Veterans Administration asymptomatic carotid stenosis study: operative morbidity and mortality. J VAsc SURG 1989;11:252-9. 3. Maini BS, Mullius TF III, Catlin J, O'Mara P. Carotid endarterectomy: a ten-year analysis of outcome and cost of treatment. I VASC SURG 1990;12:732-40. 4. Corson JD, Chang BB, Leopold PW, et al. Perioperative hypertension in patients undergoing carotid endarterectomy: shorter duration under regional block anesthesia. Circulation 1986;741:I1-4. 5. Peitzman AB, Webster MW, Loubeau JM, Grundy BL, Bahnson HT. Carotid endarterectomy under regional (conductive) anesthesia. Ann Surg 1982;196:59-64. 6. Hafner CD, Evans WE. Carotid endarterectomy with local anesthesia: results and advantages. J VAsc St3RG1988;7:232-9. 7. O'Brien MS, Ricotta JJ. Conserving resources after carotid endarterectomy: selective use of the intensive care unit. J VKsc SURG 1991;14:796-802.
Submitted May 8, 1992; accepted ~uly 1, 1992.
930
Collier
Journal of VASCULAR SURGERY
DISCUSSION
Dr. Gary Giangola (New York, N.Y.). This study demonstrates that with the use of a well-coordinated team, 88% of patients who had carotid artery operations could be discharged on the day after the procedure. By so doing, the hospital saved $1988 per patient. The question is whether this is applicable for carotid artery surgery outside of Sewicldey. Should surgeons try to establish this as a norm in their practices and will hospital administrators and third-party payers refer to this study as an encouragement for vascular surgeons in their hospitals? In evaluating _this article critically, I would first like to address the practice of outpatient angiography. This is applicable in many patients, but not in all. First, the technique of outpatient angiography usually does not include selective carotid artery injections. I find selective angiography preferable for various reasons related to the quality of the study. But in addition to quality, I would estimate that approximately 5% to 10% of the patients with carotid artery lesions at New York University (NYU) have impaired renal function and thus benefit from hydration before and after the study induced osmotic diuresis and close postangiographic monitoring of the urine output. I noticed in this series that no patients were excluded because of renal insufficiency~ so my first question to Dr, Collier concerns management of the patient with renal insuffidency. Do you admit patients with renal compromise to the hospital for angiography and if not, what precautions do you take to prevent kidney failure and how do you monitor them after the study? Blood pressure fluctuations are common after carotid endarterectomy. The author used the ICU selectively to manage hypertension. At NYU we almost never see uncontrollable hypertension as a result of limiting our dissection in the carotid bifurcation. We do see hypotension, however. Thirty percent of our patients require vasopressor support to maintain the blood pressure above a systolic level of 110 mm Hg, These patients are not placed in the ICU but rather in a postoperative unit on a nursing floor. The vasopressor support is continued for an average of 1 day after the operation, and so about a third of our patients at NYU would be unable to be discharged on the first postoperative day because of blood pressure problems alone.
Aside from blood pressure concerns, I am not sure that I would feel secure in discharging patients the day after operation. Several years ago we published data on intracerebral hemorrhage. Eleven of 1500 patients, or 0.7%, had this complication and one half of them had to be returned to the operating room for neurosurgical evacuation. The average time to occurrence of the hemorrhage was 3 V~days after operation. In addition to this complication, we recently reported the cases of three patients who had rupture of the ankle saphenous vein patch. These occurred from 2 to 5 days
postoperatively. All were fornmately still in the hospital and they were returned to the operating room immediately and had subsequent uneventful recoveries. In the April JOURNAL OF VASCULAR SURGERY, Dr. McCready from Indianapolis reported on three patients who bled after operation from needle holes in a prosthetic patch. These occurred anywhere from 1V2 to 4 days after operation. One patient had significant tracheal compression. All were returned to the operating room immediately. At NYU we have also seen several patients in whom the Prolene sutures (Ethicon, Inc.~ Somerville, N.J.) broke several days after operation. Immediate evacuation of the hematoma saved some of their lives. I would like to emphasize that these complications are not those that occur in one of every 50 patients. The incidence of these complications is less t h a n - a n d in some cases much less t h a n - 1%. So my second question for Dr. Collier is this: Would you change your practice if, say, your seventy-fifth patient or your one hundred seventy-fifth patient had a fatal complication at home on postoperative day 2 or 3? Beyond the need for blood pressure support, and in addition to the infrequent but potentially fatal complications that occur several days after carotid artery operation, is the consideration of the comfort of our patients. Most of my patients would not leave the hospital the day after this type of operation, because of their age or other circumstances, even though all procedures are performed with the patient under cervical block anesthesia. Others would find the necessary repetitive trips back to the city prohibitive. My final question to Dr. Collier involves the number of times patients have to return to his office in the first several days for common wound problems such as a drainage and edema or for pain. Finally, I would like to reemphasize that carotid artery surgical risks are measured in fractions of percentage points. Small errors, be they technical or judgmental, can result in markedly different results. Our goal should not be to see what can be done in a small group of, say, 50 patients, but it should be to achieve the best possible results in such a manner that the errors that occur but once in 200 or 300 patients are avoided. Although these results reported today are commendable, we must interpret them very cautiously because the size of the group is small. Dr. Paul E. Collier. I believe that this protocol is applicable outside of Sewickley. I think, however, that in a place like NYU or Montefiore, where I received my training, it would be difficult to perform this type of work because it is difficult for the patients to come into the city, the families are more broken up, and follow~up is harder. The first of the numerous questions had to do with arteriography. We use a direct carotid artery injection infrequently. We find that with a good duplex scan before
Volume 16 Number 6 December 1992
operation, digital subtraction arteriography can provide good results about 90% of the time. When we do not get a good picture, we do use selective carotid artery injections. In this series one patient had severe renal problems. Our protocol is to admit these patients to the hospital, begin hydratation therapy, give them mannitol, and get them in good shape medically before we do the arteriogram. As for our outpatient arteriograms, these patients come in a few hours ahead of time, are given about a liter of saline for hydration, and then have the arteriogram performed with as little dye as possible. We have not had any problems with renal failure in the patients who have had arteriography as outpatients. Hypotension is an interesting thing because we have started to notice more and more of that. In this series, two of the patients had hypotension. One was monitored in the ICU and one was monitored on a telemetry floor, and we found that even with vasopressors, it was difficult to bring their pressures up. In most of these patients, the blood pressure hovers around 100 mm H g and then gradually comes up to normal. And, at least in my experience, we really have not had any problems with patients having neurologic complications with a slightly lower blood pressure than usual. Intracerebral hemorrhage is a rare complication that has not been associated with hypertension, anticoagulation, or anything else. It just appears that when a patient has a severe stenosis that is relieved, there is a very small but real percentage of patients who are prone to intracerebral hemorrhage. In February, which was about 3 or 4 months after we finished the study, we did have a patient have an intracerebral bleed. That was the first time that occurred in one of my patients. Fortunately, for some reason, there was something about the case that made me feel uncomfortable. I didn't send the patient home the first postoperative day. The second day she looked good except she said she was tired from being in the ICU all night where she did not get any sleep. We decided to watch her a couple of more hours. Her legs got weak, and scan results showed that she had two sites of bleeding in her head. Fortunately, just with close monitoring this patient had near total resolution of symptoms without reoperation. So in answer to your question, yes, we have had one case of intracerebral hemorrhage. It does startle you, and it has made us think, but we are still sending people home on the first postoperative day. I can tell you something funny that may reflect more on my patients than on anything else. The last few patients on whom I performed carotid endarterectomy were in the hallway fully dressed at 7:10 AM waiting for me to come in to discharge them before I operated at 7:30. So they are anxious to go home. Again, this may reflect more on the type of patients I deal with and their families. Western Pennsylvania has a strong stock and very proud people who like to take care of themselves. As to the patch rupture, I am very selective with patching. I do it infrequently. In this series I did one patch,
Carotid endarterectomy 931
which was a polytetrafluoroethylene patch done during a reoperation. I find I'm patching slightly more frequently on smaller arteries, but I have not noticed the problem of rupture. I think the Cleveland Clinic recently published concerning patch rupture, and half of them occurred on the first postoperative day. The incidence was very low. Prolene suture fracture I have never seen, though I have heard about it. A friend of mine is going to court concerning a fracture in the midsuture line. Again, this problems seems to always happens about the third day. What would I do if my seventy-fifth or one hundred seventy-fifth patient had a major complication? We have had that major complication. I still think that if the patients are well selected, they may not all get to go home on the first postoperative day, but I think the majority will be able to go home early. The patients are comfortable. For some reason, I think the cervical block anesthesia provides prolonged pain relief in the neck, as opposed to general anesthesia with which patients wake up in the recovery room and require intramuscular shots for pain immediately. I think, again, this may reflect on the type of patients I deal with. Not to overbear it, but they are hardy stock and I think Fred Jarrett can probably attest to that. There are a lot of old steelworkers and coal miners in western Pennsylvania and they are a tough breed. As for returns to the office, only one patient came to the office on the third postoperative day. He had a big bull neck and his family thought that his neck was more swollen than when he went home. It was, in fact, about the same size as when he went home. I think it was just that the patient's family had not seen his neck and was uncomfortable with its appearance. This is a highly select group, and you are right, it is a small series. Unfortunately, in a small hospital we do not get the number of patients that you see, but it was experimental just to test the water and the feasibility of doing it. I think our patients have done very well and appreciated going home. Dr. Joseph Hill (Leominster, Mass.). t also practice in community hospitals. We have been using cervical block anesthesia for our patients now for 5 years. Our series is much the same as yours, and when we started using cervical block anesthesia, we immediately noticed that the bills went down, as far as hospital charges. We began looking at this, and found it was because the patients did not need the amount of intensive care monitoring and also did not need the vasopressor drugs for either hypotension or hypertension. We have, in the last 100 cases, sent the patient home routinely on the first day, rarely on the second, and have not had a major problem. Our patients also say that they want to go home. I think it can be safely done. You have to adapt the method, of course, to where you practice and to what your patients can do. If carefully done, there is no question that it can be done safely and that money can be saved. Dr. Robert A. Brigham (West Reading, Pa.). I think we have to once again look at the main thrust of this paper,
932
Collier
and that is the cost-efficiency. Having seen the program ahead of time, I asked our chief of surgery to give us some data concerning this, and very interestingly, he was able to produce for me a list of costs and the LOS for given hospitals throughout Pennsylvania for each given DRG, which the state publishes each year. He was very willing to give me the sheet for D R G 005, carotid endarterectomy, because our hospital w a s - at least of the hospitals listedthe cheapest for carotid endarterectomy in that year, 1990. It is interesting to note that the LOS was almost equal for the given study hospitals. The cost disparity, however, was significant. It ranged from $6000 at our hospital to $37,000 for a Philadelphia hospital. And again, this was all with fairly equal LOSs, ranging anywhere from 5.8 days up to 9 days. So I think that if we are looking at using this technique to reduce costs, we need to look at our individual hospital charges for things and see where we can reduce costs elsewhere, and it may not be just LOS. Dr. Paul M. Orecchia (Lemoyne, Pa.). I just have two questions. One, what happens if the patient has a temperature of, say, 100 ° F on the first postoperative day? Does that keep them in the hospital or do you send them home? And second, on your outpatient arteriogram, you demonstrated a critical carotid stenosis. In my practice, we would probably have treated the patient with heparin and performed the operation the next day. If the patient is in an outpatient stares at that point, that sort of changes the picture. What do you do? Dr. John J. Ricotta (Buffalo, N.Y.). I would like to make two comments. The key here is that you defined your indications very clearly. If that is not clone hospital, administrators or insurance agents will insist as the standard of care that you bring patients in the same day and send them home on the first postoperative day. So although I think you can do what you have done, and I commend you, I would suggest to you that we all need to define very clearly in our own hospitals what kinds of situations will allow us to send these patients home early. It should be clear that there are situations in which these people need to stay, and that needs to be indicated. My second comment concerns what Gary Giangola said about reperfusion. We found that a lot of these patients complained of a headache after operation. If my patient has a headache and any hypertension, for instance, exceeding 140 or 1 5 0 m m Hg, the patient will stay in the hospital until the headache is gone and tmtil the blood pressure is controlled, I would suggest to you that if you are going to send patients home early, you be very careful and consider this kind of symptom, which is suggestive of hyperperfusion in these patients, as a contraindication to sending a patient home early. Dr. Thomas I. O'DonneU (Boston, Mass.). This is such a provocative paper. I think one danger of this paper is that third-party payers will seize on it as the standard of care. Therefore they will urge that all patients be treated in this manner. Then the problem is, as you know, hospital administrators do not discern patient indications.
Journal of VASCULAR SURGERY
The other aspect of this paper provokes the question, "Are we really in the practice of medicine to save money for hospitals?" Instead, should we not promote the best and most cost-effective health care possible? The "real" hospital cost of carotid endarterectomy was examined in a study done 3 years ago by Massachusetts General, Brigham & Women's, and our institution in which the component cost of several years of carotid endarterectomies performed at these three hospitals was examined. As one of the previous speakers said, LOS was a very minor cost factor. The major factor driving hospital costs fbr carotid endarterectomy was arteriography, length of time in the operating room, and whether the I C U was used. Therefore, if you really want to impact hospital costs, you have to do your surgery fast, employ outpatient arteriography on a selective study, and be very selective in the use of the ICU. I think LOS is a minor factor, and as are the other discussants, I am concerned about complications being missed in short-stay surgery carotid endarterectomy. Dr. Fredric Jarrett (Pittsburgh, Pa.). I think if we are aiming at recognizing postoperative complications on the fourth or fifth day, we are probably going to miss them anyway because most everyone was sending these patients home by the third day, even before Dr. Collier's study. The other issue that has been raised is that of cervical block, and there is no question that a general anesthetic will set a patient back a little more than a regional anesthetic will. Most patients do get up more quickly and are a bit more eager and able to be discharged after a cervical block; but conversely, the method is not applicable to all patients. There are some patients who cannot lie still on the operating room table for an hour and 20 minutes or however long it takes to do the operation, and there are others for whom cervical block is not applicable. Another factor that is sometimes useful, in my experience, in getting patients out of the hospital early is not to use too many parenteral narcotics. When they are given Demerol or morphine, particularly within 24 hours of a general anesthetic, patients are groggy and sleepier and slower to come around. We have had a tremendous surge in enthusiasm in our institution for the use of ketorolac (Toradol), which is an intramuscular nonsteroidal antiinflammatory drug that has been tremendously well accepted by both the physicians and the patients. This medication often will obviate the use of a narcotic in the first 24 to 48 hours after an operation, which can be helpfial in discharging patients early if you so desire. Dr. Collier. I will attempt to answer these numerous questions. I would like to thank Dr. Hill and Dr. Jarrett for their supportive comments. I think a few things need explanation. Number one was a study by lohn Corson's group that compared cervical block with general anesthesia: their length of hypertension after cervical block averaged only 1.4 hours versus almost 20 hours for general anesthesia, which I think is another support for cervical block. I also appreciate Dr. Ricotta's comments. I think he
Volume 16 Number 6 December 1992
gave a talk here last year showing that only 18% o f his patients spent time in the I C U and that that factor alone decreased the cost o f the stay by 12.5%. There are numerous things that go into reducing costs; the LOS is a small part o f it but does contribute. I agree that the outpatient arteriography shifts a big expense to a different payment schedule, which eliminates it from the DRG. The speed of the operation is critical. I think that we are fortunate. We have a very good team that works well together. I do confess I work with one of my partners when I do the operation. I do not work with a resident so a resident does not slow me down. My partner knows my moves and I do not have to tell him how to do it, so that cuts time off the case. And as Dr. Ricotta said, the ICU, which we refer to as "expensive care," is a very important part. It is important to lower that usage. Dr. Brigham, as to the differences in cost in Pennsylvania, it is amazing the disparity is that great, although it is possible. And again I think your points are well taken. There are numerous things that go into cost. We addressed the I C U issue and we had our usage rate down to 10% which I think is a tremendous cost savings. Even the length of time in the recovery room, which bills, I think, in 15-minute units, is important. We used to keep the patients in the recovery room for 6 to 8 hours. We now have it down to 4 to 6 hours, and the head nurse in the recovery room has made me a graph that shows the drop in costs just for the recovery room time. As for the patients who have low-grade temperatures, fortunately, in Pennsylvania, we have a very strict peer review organization that has guidelines that we adhere to very strictly or we would be answering a lot o f letters. A temperature of 100 ° F, to me, is just a low-grade temperature. I do not get too excited about it. We do not use Foley catheters in our patients undergoing carotid procedures so I do not worry about the urine. I use 101 ° F as a sort of a cutoff. I think that many people after operation have a temperature that goes to 100 ° F. At 101 ° F I would get worried. I would probably keep the patients in the hospital, and I would check their urine. But carotid endarterectomies are nice procedures. The wounds rarely get infected. In 6 years I have seen one wound infection. The patient had chronic external otitis, which drained right next to the wound, and the wound got a superficial infection. Other than that, I have not seen an infection with this procedure. I think that most o f us have experienced an extremely low rate of infection in the neck. Again, if a Foley catheter is not being used, the patient's chance o f getting a urinary problem is not significant. About the only thing you need to worry about is that if the larynx is numbed with the cervical block the patient may have a problem with some minor aspiration and you might need to check the chest radiograph. As for critical stenoses, this is something that I have
Carotid endarterectomy 933
dealt with. I think that in the literature the first mention of the use of heparin preoperatively was by Dr. Wes Moore. I was curious about his report because I had seen a couple of mesenteric thromboses after arteriography and I talked to him to find out where he got the literature references. He said it was just his gut feeling that heparin should be used and that that was the way he handled these lesions. H e said because it was just an article in his review book he did not have to document it anywhere. But when I looked through the literature, there was really no good documentation that patients with severe stenoses are prone to sudden thromboses if they are not heparinized. We still closely watch the patients with 99% stenoses, and we operate on them as expeditiously as possible. And we are nervous, but I have not found that bringing them into the hospital and giving them heparin prevents anything from happening. I was with a friend of mine last night, Dr. Steve Rivers, who follows that procedure. H e asked me about this question. He told me about one patient he had admitted to the hospital and had placed on heparin overnight. On the morning of operation she had had a stroke despite having received heparin. I have not found that using heparin for severe stenoses does anything to prevent any complications. I agree with Dr. Ricotta and Dr. O'Donnell that we need to be very strict with our indications. The reperfusion problem is something that scares me, especially now that I have seen a postoperative intracerebral bleed. Subtle things made me keep the patient in the hospital. She had a minor headache. It was not the type of headache that you get too excited about. She was also slightly nauseated. We naturally think back to our training and to patients who come for treatment with intracerebrai hemorrhages: there is usually a headache that is the worst of their life, a little nausea, maybe a little vomiting, and as they get worse, neurologic symptoms develop. In this series anyone who had anything that made me nervous was kept in the hospital, and these things included headaches, nausea, and vomiting. If the patient could not hold down their food, I did not send the patient home. And I think that the LOS, although it is a small contributor, is an overall contributor. I hope the third-party payers will read all the comments after the talk and realize that this is a select group o f patients. Early discharge is something that we all aim for. We are not necessarily trying to save money for the hospital, but we are from Pennsylvania, where we have Senator Wofford who is out there campaigning for national health insurance and saying that doctors are the ones who run up the charges. I think that if we physicians show an active interest in keeping the charges down, as Dr. Ricotta did in his study last year, perhaps the legislators will realize that we are actually on the same team and work together more with Dr. Veith's group to make medicine work in our country.
The introduction of the diagnosis-related group (DRG) system challenged hospitals and physicians to scrutinize their efficiency in delivering services. One of the most dramatic results of prospective payment has been the gradual reduction in the hospital length of stay (LOS) for many surgical procedures. This has also led to a reduction in the cost of delivering these same services. Studies such as the North American Symptomatic Carotid Endarterectomy Trial and Veterans Administration cooperative studies have shown that carotid endarterectomy can be performed safely in experienced hands. 1'2 However, questions remain about From the Department of Surgery, SewickleyValley Hospital, Sewickley. Presented at the Sixth Annual Meeting of the Eastern Vascular Society,New York, N.Y., April 30-May3, 1992. Reprint requests: PaulE. Collier,MD, Suite200, 301 Ohio River Blvd., Sewickley,PA 15143. 24/6/40729 926
what an appropriate LOS for a patient after a carotid endarterectomy should be and whether there will be an increase in morbidity or mortality with earlier discharge. Originally patients tentatively admitted to the hospital for carotid endarterectomy at this institution were admitted the clay before operation for arteriography unless they were already in the hospital. If the arteriogram showed an operable lesion, the patient underwent a carotid endarterectomy the following day. The majority of patients were operated on with the use of cervical block anesthesia and selective shunting. Postoperatively, all patients were monitored in the intensive care unit (ICU) for 24 hours and then discharged on the second or third postoperative day. After reviewing the hospital records of all patients who underwent carotid endarterectomy between 1988 and 1990 and directly interviewing many of these patients, the following hypotheses were made.
Volume 16 Number 6 December 1992
Arteriography could safely be performed on an outpatient basis before carotid endarterectomy. Patients undergoing carotid endarterectomy could be admitted to the hospital on the day of operation and the use of the ICU postoperatively could be limited to select patients. Finally, it was hypothesized that patients could be discharged early, often on the first postoperative day, without any adverse effects. A protocol was established to prospectively evaluate these hypotheses. The results of that study are the basis of this report.
PATIENTS AND METHODS Between January 1 and October 31, 1991, 51 patients underwent 52 carotid endarterectomies at Sewickley Valley Hospital. There were 36 men and 15 women ranging in age from 44 to 84 years. All 51 patients were entered in the following protocol. Unless the patient was in the hospital, a history, physical examination, and review of the duplex scan were performed in the office. If the patient was thought to be a suitable candidate for carotid endarterectomy, an outpatient arteriogram was scheduled. During this initial patient encounter the technique and possible complications of arteriography were discussed with the patient. More important, the feasibility of outpatient arteriography and the safety record of the radiology department were stressed so that the patient would be comfortable with the outpatient procedure. If the arteriogram demonstrated an operable lesion, the patient was scheduled for operation when he or she was next seen in the office. At that time, the risks and possible complications of the operation were discussed with the patient, and whenever feasible, a family member was present during this discussion. During this encounter our results and the possibility of early discharge were discussed with the patient so he or she could make arrangements ahead of time and feel comfortable at the time of discharge. Whenever possible, cervical block was used for anesthesia and shunting was done selectively. Patients were closely monitored in the recovery room for a minimum of 4 hours. Blood pressure was monitored with indwelling arterial lines in all patients. If the patients were unchanged neurologically, had no cardiac problems, and maintained a systolic blood pressure less than 180 mm H g without medication, they were transferred to the vascular floor; otherwise, they were monitored in the ICU. Patients were discharged on the first postoperative day if their neurologic status was unchanged, there were no problems with their blood pressure or heart, their wound was satisfactory,
Carotid endarterectomy 927
and they felt comfortable going home at that time. All patients were seen in the office between 3 and 5 days postoperatively. Of the 51 patients undergoing carotid endarterectomy, 26 (51%) were receiving antihypertensive medication, seven (14%) were diabetic, and 23 (45%) were smokers. Twenty-seven (53%) had coronary artery disease manifested by either angina pectoris, a history of myocardial infarction, or a prior coronary artery bypass. Thirty-one (60%) of the operations were performed for transient ischemic attacks or amaurosis fugax, 11 (21%) for a prior stroke, and 10 (19%) for an asymptomatic stenosis greater than 80%. Forty-three carotid endarterectomies were performed with cervical block anesthesia. Eight (19%) of these patients required shunts. Nine patients required general anesthesia and all of these had shunts. The reasons for general anesthesia included recent stroke, patient anxiety, and one late reoperation for a recurrent, symptomatic stenosis. RESULTS Three arteriograms were performed on patients who were already in the hospital. Of the 48 other arteriograms all were successfully performed as outpatient procedures. The only complication encountered was a small groin hematoma, which resolved with conservative therapy. There were no neurologic, cardiac, or renal complications. None of the 48 patients required admission to the hospital as a result of the angiogram. Two neurologic events occurred as a result of the 52 carotid endarterectomies. One patient had weakness of the left arm that lasted less than 5 minutes while the patient was in the recovery room; there was subsequent total recovery. Duplex scanning demonstrated a patent repair with no defects. A second patient had undergone carotid endarterectomy after a stroke with near total resolution 6 weeks previously. After she had vomited in the recovery room, bradycardia and hypotension developed with recurrence of the prior stroke symptoms. Restoring her blood pressure did not change her symptoms so she was rapidly returned to the operating room. No thrombus or technical defect was found at reexploration. She was left with residual symptoms, worse than she had preoperatively. She now functions independently with minimal residual symptoms. For this group of 52 operations, the permanent stroke rate was 1.9%. Both neurologic events occurred within 2 hours after the operation. No neurologic events occurred after the patients were discharged from the recovery room. Transient problems with hypertension were corn-
Journal of VASCULAR SURGERY
928 Collier
mon. Twenty-eight patients (54%) required continuous intravenous nitroglycerin administration to maintain a systolic blood pressure less than 180 mm Hg intraoperatively. Twenty-six patients (50%) required antihypertensive treatment in the recovery room. Of the 24 patients whose blood pressure was controlled in the recovery room, 23 (96%) had therapy terminated within 2 hours. The other patient's treatment lasted 3 hours. Only two patients (4%) did not have their blood pressure normalized while in the recovery room and required further treatment and monitoring in the ICU. Antihypertensive therapy was required for 8 and 9 hours in these two patients. Persistent hypotension developed in two patients (systolic blood pressure less than 100 mm Hg). One was monitored in the ICU overnight and the other was monitored on a telemetry floor without any adverse effects. Retrospectively, neither patient required active ICU treatment. Only five of the 52 patients (10%) required postoperative monitoring in the ICU: two for neurologic changes, two for persistent hypertension, and one for persistent hypotension. No patient remained in the ICU for more than 24 hours. Forty-six patients (88%) were discharged after only 1 day in t h e hospital. Two patients were discharged on the second postoperative day, two on the fourth, and one each on the third and sixth postoperative days. The average LOS was 1.29 days per operation. The average hospital charge was $3408 per patient admission. DRG 005 (Extracranial Vascular Procedures) reimburses the hospital $5396. This resulted in an average profit to the hospital of $1988 for each carotid endarterectomy performed during the 10-month study period. There were no readmissions or emergency department visits for hypertension, hemorrhage, or cardiac or neurologic events. Because of a quirk in the Medicare rules and regulations, an outpatient arteriogram is reimbursed outside of the DRG if it is performed more than 72 hours before the hospital admission. These data were therefore retrospectively reevaluated to determine whether there was an overall cost reduction for the system following this protocol as compared with performing an inpatient arteriogram and extending the length of stay by 1 day. For Medicare patients this protocol cost the system $900 more per patient when compared with inpatient arteriography. From a social standpoint, however, this protocol provides improved bed utilization for sick patients in a hospital
that is often operating at full capacity. Because reimbursement by all other payers, like Blue Cross, is based on actual costs and charges, it was found that the protocol that used outpatient arteriography was slightly less expensive because the cost of the short stay unit was less than the cost of an extra day in the hospital. After discharge the patients were interviewed regarding their satisfaction with the protocol. None of the patients believed that they had been forced to leave the hospital before they were ready. Most patients thought they did better at home. Only three patients complained that they would have preferred to stay in the hospital after their arteriogram, but they were happy to be home after their operation. DISCUSSION
The results of this study demonstrate that the protocol used is safe for the patient and cost-efficient for the hospital. Some of this efficiency is secondary to the cooperation between the emergency physicians, primary care physicians, radiologists, and vascular surgeons all working in one institution. Patients with transient neurologic symptoms and normal computerized tomographic scans are not routinely admitted to the hospital. They undergo an expeditious outpatient workup and are then admitted if they are a candidate for operation. The shifting of the arteriogram from the inpatient to the outpatient setting accomplishes two ends. First, it removes the expense of the arteriogram from the hospital cost that is reimbursed by DRG 005, and second, it reduces the hospital LOS by 1 day. The cost analysis of carotid endarterectomy done by Maini et al.3 included the arteriogram as part of the inpatient admission. This lengthened the reported LOS and cost. In our study there were no complications in any of the selective carotid arteriograms performed. In our study digital subtraction intraarterial techniques were used to reduce the dye load, and direct carotid injections were only used if unsatisfactory pictures were obtained. Limiting the use of direct carotid injections reduces the risk of procedure-related stroke. Current technology allows the use of very small (4F or 5F) catheters for arterial cannulation, which therefore reduces the risk of hematoma. The skill of the radiologists has virtually eliminated local arterial complications. These results clearly demonstrate the safety of outpatient carotid arteriography in both patients with symptoms and those without symptoms.
Volume 16 Number 6 December 1992
Same-day hospital admission for carotid endarterectomy procedures is possible only if the patient's blood pressure is well controlled and their cardiac status is satisfactory. Corson et al.4 showed that perioperative hypertension correlated best with uncontrolled preoperative hypertension. They believed that good preoperative blood pressure control would provide better perioperative control. This is demonstrated by the current series in which all patients had well-regulated blood pressure before admission, with the result that only two patients required ICU treatment for prolonged hypertension. Cardiac status is evaluated by history and resting electrocardiogram. Dipyridamolethallium scanning is obtained only if the clinical picture or electrocardiogram warrant it. In light of the reported mortality rates of less than 1% for carotid endarterectomy with regional anesthesia,5,6 dipyridamole-thallium stress tests are reserved for patients with worrisome cardiac symptoms. The 0% myocardial infarction rate in this series again attests to the low cardiac risk for carotid endarterectomy done with regional block. Regional anesthesia was used in the majority of patients for three reasons. First, as shown by Corson et al., 4 when perioperative hypertension occurs, it is of shorter duration when regional anesthesia is used. In their series postoperative hypertension averaged 1.4 hours when regional anesthesia was used and 19.6 hours when general anesthesia was used. In the current series only two patients had hypertension lasting more than 3 hours postoperatively that therefore required ICU care. One patient had been operated on with general anesthesia and the other with regional block. The second reason for using regional anesthesia is the extremely low myocardial infarction and mortality rates reported? ,S,6 Finally, regional anesthesia is used because it provides the best method of neurologic monitoring during carotid cross-clamping. Selective use of the ICU after carotid endarterectomy has been shown to be reasonable and cost effective by O'Brien and Ricotta. 7 In their series only 18% of patients required a stay in the ICU for an average of 24.5 hours. Their study showed that those patients requiring ICU care can be identified in the early postoperative period. They showed that the charge for the ICU added 12.5% to the patient's bill. The two myocardial infarctions in their series occurred 2 hours and 6 hours postoperatively. In the current series there were no myocardial infarctions
Carotid endarterectomy
929
and both neurologic events occurred while the patients were still in the recovery room. The safety of this selective policy is demonstrated by the fact that there were no late admissions or readmissions to the ICU. Because only 10% of the patients required ICU stays in this series, 90% of the patient's hospital charges were lower by 12.5%. The LOS for most procedures including carotid endarterectomy has been reduced by the introduction of prospective payment? In this series the average LOS was reduced to 1.29 days per patient. This was done by the use of outpatient arteriography, same-day hospital admission for the procedure, regional anesthesia, and early discharge when feasible. The cost efficiency of this approach is clear. The safety is demonstrated by the fact that there were no returns to the hospital for uncontrolled hypertension, bleeding, or neurologic or cardiac problems, and there were no mortalities within 30 days of operation. Multiple factors contribute to the safety and efficiency of carotid endarterectomy. From the data presented in this study it is concluded that this protocol of regional anesthesia, selective use of the ICU, and early hospital discharge is a safe and extremely cost-effective method of treatment for patients undergoing carotid endarterectomy. REFERENCES 1. North American Symptomatic Carotid Endarterectomy Trial (NASCET): North American Symptomatic Carotid Endarterectomy Trial: methods, patient characteristics and progress. Stroke 1991;22:711-20. 2. Towne JB, Weiss DG, Hobson RW II. First phase report of cooperative Veterans Administration asymptomatic carotid stenosis study: operative morbidity and mortality. J VAsc SURG 1989;11:252-9. 3. Maini BS, Mullius TF III, Catlin J, O'Mara P. Carotid endarterectomy: a ten-year analysis of outcome and cost of treatment. I VASC SURG 1990;12:732-40. 4. Corson JD, Chang BB, Leopold PW, et al. Perioperative hypertension in patients undergoing carotid endarterectomy: shorter duration under regional block anesthesia. Circulation 1986;741:I1-4. 5. Peitzman AB, Webster MW, Loubeau JM, Grundy BL, Bahnson HT. Carotid endarterectomy under regional (conductive) anesthesia. Ann Surg 1982;196:59-64. 6. Hafner CD, Evans WE. Carotid endarterectomy with local anesthesia: results and advantages. J VAsc St3RG1988;7:232-9. 7. O'Brien MS, Ricotta JJ. Conserving resources after carotid endarterectomy: selective use of the intensive care unit. J VKsc SURG 1991;14:796-802.
Submitted May 8, 1992; accepted ~uly 1, 1992.
930
Collier
Journal of VASCULAR SURGERY
DISCUSSION
Dr. Gary Giangola (New York, N.Y.). This study demonstrates that with the use of a well-coordinated team, 88% of patients who had carotid artery operations could be discharged on the day after the procedure. By so doing, the hospital saved $1988 per patient. The question is whether this is applicable for carotid artery surgery outside of Sewicldey. Should surgeons try to establish this as a norm in their practices and will hospital administrators and third-party payers refer to this study as an encouragement for vascular surgeons in their hospitals? In evaluating _this article critically, I would first like to address the practice of outpatient angiography. This is applicable in many patients, but not in all. First, the technique of outpatient angiography usually does not include selective carotid artery injections. I find selective angiography preferable for various reasons related to the quality of the study. But in addition to quality, I would estimate that approximately 5% to 10% of the patients with carotid artery lesions at New York University (NYU) have impaired renal function and thus benefit from hydration before and after the study induced osmotic diuresis and close postangiographic monitoring of the urine output. I noticed in this series that no patients were excluded because of renal insufficiency~ so my first question to Dr, Collier concerns management of the patient with renal insuffidency. Do you admit patients with renal compromise to the hospital for angiography and if not, what precautions do you take to prevent kidney failure and how do you monitor them after the study? Blood pressure fluctuations are common after carotid endarterectomy. The author used the ICU selectively to manage hypertension. At NYU we almost never see uncontrollable hypertension as a result of limiting our dissection in the carotid bifurcation. We do see hypotension, however. Thirty percent of our patients require vasopressor support to maintain the blood pressure above a systolic level of 110 mm Hg, These patients are not placed in the ICU but rather in a postoperative unit on a nursing floor. The vasopressor support is continued for an average of 1 day after the operation, and so about a third of our patients at NYU would be unable to be discharged on the first postoperative day because of blood pressure problems alone.
Aside from blood pressure concerns, I am not sure that I would feel secure in discharging patients the day after operation. Several years ago we published data on intracerebral hemorrhage. Eleven of 1500 patients, or 0.7%, had this complication and one half of them had to be returned to the operating room for neurosurgical evacuation. The average time to occurrence of the hemorrhage was 3 V~days after operation. In addition to this complication, we recently reported the cases of three patients who had rupture of the ankle saphenous vein patch. These occurred from 2 to 5 days
postoperatively. All were fornmately still in the hospital and they were returned to the operating room immediately and had subsequent uneventful recoveries. In the April JOURNAL OF VASCULAR SURGERY, Dr. McCready from Indianapolis reported on three patients who bled after operation from needle holes in a prosthetic patch. These occurred anywhere from 1V2 to 4 days after operation. One patient had significant tracheal compression. All were returned to the operating room immediately. At NYU we have also seen several patients in whom the Prolene sutures (Ethicon, Inc.~ Somerville, N.J.) broke several days after operation. Immediate evacuation of the hematoma saved some of their lives. I would like to emphasize that these complications are not those that occur in one of every 50 patients. The incidence of these complications is less t h a n - a n d in some cases much less t h a n - 1%. So my second question for Dr. Collier is this: Would you change your practice if, say, your seventy-fifth patient or your one hundred seventy-fifth patient had a fatal complication at home on postoperative day 2 or 3? Beyond the need for blood pressure support, and in addition to the infrequent but potentially fatal complications that occur several days after carotid artery operation, is the consideration of the comfort of our patients. Most of my patients would not leave the hospital the day after this type of operation, because of their age or other circumstances, even though all procedures are performed with the patient under cervical block anesthesia. Others would find the necessary repetitive trips back to the city prohibitive. My final question to Dr. Collier involves the number of times patients have to return to his office in the first several days for common wound problems such as a drainage and edema or for pain. Finally, I would like to reemphasize that carotid artery surgical risks are measured in fractions of percentage points. Small errors, be they technical or judgmental, can result in markedly different results. Our goal should not be to see what can be done in a small group of, say, 50 patients, but it should be to achieve the best possible results in such a manner that the errors that occur but once in 200 or 300 patients are avoided. Although these results reported today are commendable, we must interpret them very cautiously because the size of the group is small. Dr. Paul E. Collier. I believe that this protocol is applicable outside of Sewickley. I think, however, that in a place like NYU or Montefiore, where I received my training, it would be difficult to perform this type of work because it is difficult for the patients to come into the city, the families are more broken up, and follow~up is harder. The first of the numerous questions had to do with arteriography. We use a direct carotid artery injection infrequently. We find that with a good duplex scan before
Volume 16 Number 6 December 1992
operation, digital subtraction arteriography can provide good results about 90% of the time. When we do not get a good picture, we do use selective carotid artery injections. In this series one patient had severe renal problems. Our protocol is to admit these patients to the hospital, begin hydratation therapy, give them mannitol, and get them in good shape medically before we do the arteriogram. As for our outpatient arteriograms, these patients come in a few hours ahead of time, are given about a liter of saline for hydration, and then have the arteriogram performed with as little dye as possible. We have not had any problems with renal failure in the patients who have had arteriography as outpatients. Hypotension is an interesting thing because we have started to notice more and more of that. In this series, two of the patients had hypotension. One was monitored in the ICU and one was monitored on a telemetry floor, and we found that even with vasopressors, it was difficult to bring their pressures up. In most of these patients, the blood pressure hovers around 100 mm H g and then gradually comes up to normal. And, at least in my experience, we really have not had any problems with patients having neurologic complications with a slightly lower blood pressure than usual. Intracerebral hemorrhage is a rare complication that has not been associated with hypertension, anticoagulation, or anything else. It just appears that when a patient has a severe stenosis that is relieved, there is a very small but real percentage of patients who are prone to intracerebral hemorrhage. In February, which was about 3 or 4 months after we finished the study, we did have a patient have an intracerebral bleed. That was the first time that occurred in one of my patients. Fortunately, for some reason, there was something about the case that made me feel uncomfortable. I didn't send the patient home the first postoperative day. The second day she looked good except she said she was tired from being in the ICU all night where she did not get any sleep. We decided to watch her a couple of more hours. Her legs got weak, and scan results showed that she had two sites of bleeding in her head. Fortunately, just with close monitoring this patient had near total resolution of symptoms without reoperation. So in answer to your question, yes, we have had one case of intracerebral hemorrhage. It does startle you, and it has made us think, but we are still sending people home on the first postoperative day. I can tell you something funny that may reflect more on my patients than on anything else. The last few patients on whom I performed carotid endarterectomy were in the hallway fully dressed at 7:10 AM waiting for me to come in to discharge them before I operated at 7:30. So they are anxious to go home. Again, this may reflect more on the type of patients I deal with and their families. Western Pennsylvania has a strong stock and very proud people who like to take care of themselves. As to the patch rupture, I am very selective with patching. I do it infrequently. In this series I did one patch,
Carotid endarterectomy 931
which was a polytetrafluoroethylene patch done during a reoperation. I find I'm patching slightly more frequently on smaller arteries, but I have not noticed the problem of rupture. I think the Cleveland Clinic recently published concerning patch rupture, and half of them occurred on the first postoperative day. The incidence was very low. Prolene suture fracture I have never seen, though I have heard about it. A friend of mine is going to court concerning a fracture in the midsuture line. Again, this problems seems to always happens about the third day. What would I do if my seventy-fifth or one hundred seventy-fifth patient had a major complication? We have had that major complication. I still think that if the patients are well selected, they may not all get to go home on the first postoperative day, but I think the majority will be able to go home early. The patients are comfortable. For some reason, I think the cervical block anesthesia provides prolonged pain relief in the neck, as opposed to general anesthesia with which patients wake up in the recovery room and require intramuscular shots for pain immediately. I think, again, this may reflect on the type of patients I deal with. Not to overbear it, but they are hardy stock and I think Fred Jarrett can probably attest to that. There are a lot of old steelworkers and coal miners in western Pennsylvania and they are a tough breed. As for returns to the office, only one patient came to the office on the third postoperative day. He had a big bull neck and his family thought that his neck was more swollen than when he went home. It was, in fact, about the same size as when he went home. I think it was just that the patient's family had not seen his neck and was uncomfortable with its appearance. This is a highly select group, and you are right, it is a small series. Unfortunately, in a small hospital we do not get the number of patients that you see, but it was experimental just to test the water and the feasibility of doing it. I think our patients have done very well and appreciated going home. Dr. Joseph Hill (Leominster, Mass.). t also practice in community hospitals. We have been using cervical block anesthesia for our patients now for 5 years. Our series is much the same as yours, and when we started using cervical block anesthesia, we immediately noticed that the bills went down, as far as hospital charges. We began looking at this, and found it was because the patients did not need the amount of intensive care monitoring and also did not need the vasopressor drugs for either hypotension or hypertension. We have, in the last 100 cases, sent the patient home routinely on the first day, rarely on the second, and have not had a major problem. Our patients also say that they want to go home. I think it can be safely done. You have to adapt the method, of course, to where you practice and to what your patients can do. If carefully done, there is no question that it can be done safely and that money can be saved. Dr. Robert A. Brigham (West Reading, Pa.). I think we have to once again look at the main thrust of this paper,
932
Collier
and that is the cost-efficiency. Having seen the program ahead of time, I asked our chief of surgery to give us some data concerning this, and very interestingly, he was able to produce for me a list of costs and the LOS for given hospitals throughout Pennsylvania for each given DRG, which the state publishes each year. He was very willing to give me the sheet for D R G 005, carotid endarterectomy, because our hospital w a s - at least of the hospitals listedthe cheapest for carotid endarterectomy in that year, 1990. It is interesting to note that the LOS was almost equal for the given study hospitals. The cost disparity, however, was significant. It ranged from $6000 at our hospital to $37,000 for a Philadelphia hospital. And again, this was all with fairly equal LOSs, ranging anywhere from 5.8 days up to 9 days. So I think that if we are looking at using this technique to reduce costs, we need to look at our individual hospital charges for things and see where we can reduce costs elsewhere, and it may not be just LOS. Dr. Paul M. Orecchia (Lemoyne, Pa.). I just have two questions. One, what happens if the patient has a temperature of, say, 100 ° F on the first postoperative day? Does that keep them in the hospital or do you send them home? And second, on your outpatient arteriogram, you demonstrated a critical carotid stenosis. In my practice, we would probably have treated the patient with heparin and performed the operation the next day. If the patient is in an outpatient stares at that point, that sort of changes the picture. What do you do? Dr. John J. Ricotta (Buffalo, N.Y.). I would like to make two comments. The key here is that you defined your indications very clearly. If that is not clone hospital, administrators or insurance agents will insist as the standard of care that you bring patients in the same day and send them home on the first postoperative day. So although I think you can do what you have done, and I commend you, I would suggest to you that we all need to define very clearly in our own hospitals what kinds of situations will allow us to send these patients home early. It should be clear that there are situations in which these people need to stay, and that needs to be indicated. My second comment concerns what Gary Giangola said about reperfusion. We found that a lot of these patients complained of a headache after operation. If my patient has a headache and any hypertension, for instance, exceeding 140 or 1 5 0 m m Hg, the patient will stay in the hospital until the headache is gone and tmtil the blood pressure is controlled, I would suggest to you that if you are going to send patients home early, you be very careful and consider this kind of symptom, which is suggestive of hyperperfusion in these patients, as a contraindication to sending a patient home early. Dr. Thomas I. O'DonneU (Boston, Mass.). This is such a provocative paper. I think one danger of this paper is that third-party payers will seize on it as the standard of care. Therefore they will urge that all patients be treated in this manner. Then the problem is, as you know, hospital administrators do not discern patient indications.
Journal of VASCULAR SURGERY
The other aspect of this paper provokes the question, "Are we really in the practice of medicine to save money for hospitals?" Instead, should we not promote the best and most cost-effective health care possible? The "real" hospital cost of carotid endarterectomy was examined in a study done 3 years ago by Massachusetts General, Brigham & Women's, and our institution in which the component cost of several years of carotid endarterectomies performed at these three hospitals was examined. As one of the previous speakers said, LOS was a very minor cost factor. The major factor driving hospital costs fbr carotid endarterectomy was arteriography, length of time in the operating room, and whether the I C U was used. Therefore, if you really want to impact hospital costs, you have to do your surgery fast, employ outpatient arteriography on a selective study, and be very selective in the use of the ICU. I think LOS is a minor factor, and as are the other discussants, I am concerned about complications being missed in short-stay surgery carotid endarterectomy. Dr. Fredric Jarrett (Pittsburgh, Pa.). I think if we are aiming at recognizing postoperative complications on the fourth or fifth day, we are probably going to miss them anyway because most everyone was sending these patients home by the third day, even before Dr. Collier's study. The other issue that has been raised is that of cervical block, and there is no question that a general anesthetic will set a patient back a little more than a regional anesthetic will. Most patients do get up more quickly and are a bit more eager and able to be discharged after a cervical block; but conversely, the method is not applicable to all patients. There are some patients who cannot lie still on the operating room table for an hour and 20 minutes or however long it takes to do the operation, and there are others for whom cervical block is not applicable. Another factor that is sometimes useful, in my experience, in getting patients out of the hospital early is not to use too many parenteral narcotics. When they are given Demerol or morphine, particularly within 24 hours of a general anesthetic, patients are groggy and sleepier and slower to come around. We have had a tremendous surge in enthusiasm in our institution for the use of ketorolac (Toradol), which is an intramuscular nonsteroidal antiinflammatory drug that has been tremendously well accepted by both the physicians and the patients. This medication often will obviate the use of a narcotic in the first 24 to 48 hours after an operation, which can be helpfial in discharging patients early if you so desire. Dr. Collier. I will attempt to answer these numerous questions. I would like to thank Dr. Hill and Dr. Jarrett for their supportive comments. I think a few things need explanation. Number one was a study by lohn Corson's group that compared cervical block with general anesthesia: their length of hypertension after cervical block averaged only 1.4 hours versus almost 20 hours for general anesthesia, which I think is another support for cervical block. I also appreciate Dr. Ricotta's comments. I think he
Volume 16 Number 6 December 1992
gave a talk here last year showing that only 18% o f his patients spent time in the I C U and that that factor alone decreased the cost o f the stay by 12.5%. There are numerous things that go into reducing costs; the LOS is a small part o f it but does contribute. I agree that the outpatient arteriography shifts a big expense to a different payment schedule, which eliminates it from the DRG. The speed of the operation is critical. I think that we are fortunate. We have a very good team that works well together. I do confess I work with one of my partners when I do the operation. I do not work with a resident so a resident does not slow me down. My partner knows my moves and I do not have to tell him how to do it, so that cuts time off the case. And as Dr. Ricotta said, the ICU, which we refer to as "expensive care," is a very important part. It is important to lower that usage. Dr. Brigham, as to the differences in cost in Pennsylvania, it is amazing the disparity is that great, although it is possible. And again I think your points are well taken. There are numerous things that go into cost. We addressed the I C U issue and we had our usage rate down to 10% which I think is a tremendous cost savings. Even the length of time in the recovery room, which bills, I think, in 15-minute units, is important. We used to keep the patients in the recovery room for 6 to 8 hours. We now have it down to 4 to 6 hours, and the head nurse in the recovery room has made me a graph that shows the drop in costs just for the recovery room time. As for the patients who have low-grade temperatures, fortunately, in Pennsylvania, we have a very strict peer review organization that has guidelines that we adhere to very strictly or we would be answering a lot o f letters. A temperature of 100 ° F, to me, is just a low-grade temperature. I do not get too excited about it. We do not use Foley catheters in our patients undergoing carotid procedures so I do not worry about the urine. I use 101 ° F as a sort of a cutoff. I think that many people after operation have a temperature that goes to 100 ° F. At 101 ° F I would get worried. I would probably keep the patients in the hospital, and I would check their urine. But carotid endarterectomies are nice procedures. The wounds rarely get infected. In 6 years I have seen one wound infection. The patient had chronic external otitis, which drained right next to the wound, and the wound got a superficial infection. Other than that, I have not seen an infection with this procedure. I think that most o f us have experienced an extremely low rate of infection in the neck. Again, if a Foley catheter is not being used, the patient's chance o f getting a urinary problem is not significant. About the only thing you need to worry about is that if the larynx is numbed with the cervical block the patient may have a problem with some minor aspiration and you might need to check the chest radiograph. As for critical stenoses, this is something that I have
Carotid endarterectomy 933
dealt with. I think that in the literature the first mention of the use of heparin preoperatively was by Dr. Wes Moore. I was curious about his report because I had seen a couple of mesenteric thromboses after arteriography and I talked to him to find out where he got the literature references. He said it was just his gut feeling that heparin should be used and that that was the way he handled these lesions. H e said because it was just an article in his review book he did not have to document it anywhere. But when I looked through the literature, there was really no good documentation that patients with severe stenoses are prone to sudden thromboses if they are not heparinized. We still closely watch the patients with 99% stenoses, and we operate on them as expeditiously as possible. And we are nervous, but I have not found that bringing them into the hospital and giving them heparin prevents anything from happening. I was with a friend of mine last night, Dr. Steve Rivers, who follows that procedure. H e asked me about this question. He told me about one patient he had admitted to the hospital and had placed on heparin overnight. On the morning of operation she had had a stroke despite having received heparin. I have not found that using heparin for severe stenoses does anything to prevent any complications. I agree with Dr. Ricotta and Dr. O'Donnell that we need to be very strict with our indications. The reperfusion problem is something that scares me, especially now that I have seen a postoperative intracerebral bleed. Subtle things made me keep the patient in the hospital. She had a minor headache. It was not the type of headache that you get too excited about. She was also slightly nauseated. We naturally think back to our training and to patients who come for treatment with intracerebrai hemorrhages: there is usually a headache that is the worst of their life, a little nausea, maybe a little vomiting, and as they get worse, neurologic symptoms develop. In this series anyone who had anything that made me nervous was kept in the hospital, and these things included headaches, nausea, and vomiting. If the patient could not hold down their food, I did not send the patient home. And I think that the LOS, although it is a small contributor, is an overall contributor. I hope the third-party payers will read all the comments after the talk and realize that this is a select group o f patients. Early discharge is something that we all aim for. We are not necessarily trying to save money for the hospital, but we are from Pennsylvania, where we have Senator Wofford who is out there campaigning for national health insurance and saying that doctors are the ones who run up the charges. I think that if we physicians show an active interest in keeping the charges down, as Dr. Ricotta did in his study last year, perhaps the legislators will realize that we are actually on the same team and work together more with Dr. Veith's group to make medicine work in our country.