The influence of surgical specialty and caseload on the results of carotid endarterectomy

Transcription

1 The influence of surgical specialty and caseload on the results of carotid endarterectomy Richard F. Kempczinski, M.D., Thomas G. Brott, M.D., and Robert J. Labutta, M.D., Cincinnati, Ohio Carotid endarterectomy is rapidly becoming one of the most commonly performed major surgical operations in the United States, in part because of the greater availability of noninvasive techniques to accurately diagnose extracranial carotid arterial disease and a low reported morbidity and mortality. We retrospectively reviewed the records for all carotid endarterectomies performed in the greater Cincinnati area for a recent 12-month period and examined the impact of surgical specialty and operative caseload on the results. Altogether, 750 operations were Performed on 656 patients by 61 surgeons working in 16 general medical,surgical hospitals. Overall, strokes occurred in 5.1% of all patients; 2.3% of patients died. Symptomatic patients had a significantly higher risk of suffering a postoperative stroke compared with asymptomatic patients (6.5% vs. 3.7%), although the risk of death was virtually identical (2.4% vs. 2.1%). When the operating surgeons were classified into four types on the basis of their previous training, no statistically significant differences in either postoperative stroke or death could be identified. Furthermore, when the surgical caseloads of these physicians were grouped into three categories (i.e., less than 12 each year, more than 50 each year, and a group between these two extremes), no significant differences in outcome were seen. We concluded that our community-wide results for carotid endarterectomy were not comparable to those previously published from specialized centers and that these results did not appear to be influenced by the type of formal surgical specialty or operative caseload. (J VAse SURG 1986; 3:911-6.) Although the exact number of carotid endarterectomies performed annually in the United States is unknown, at least one report now lists it as the most common, noncardiac, major vascular operation rccorded.~ Reflecting this growth, a recent nationwide study documented a fivefold increase in the number of carotid cndarterectomies performed during the past 10 years? Several factors appear responsible for this tremendous increase in the frequency of carotid endarterectomy, including the widespread availability of noninvasive cerebrovascular testing, especially realtime B-mode imaging; easy access to digital subtraction angiography; and a greater acceptance by the medical community of the place of carotid endarterectomy in the treatment of asymptomatic carotid stenosis and, more recently, asymptomatic carotid ul- From the Division of Vascular Surgery, Department of Surgery, and the Department of Neurology, University of Cincinnati Medical Center. Reprint requests: Richard F. Kempczinski, M.D., Department of Surgery, University of Cincinnati Medical Center, 231 Bethesda Ave., Cincinnati, OH ceration? Part of the justification for operating on asymptomatic patients is based on a purportedly low operative morbidity and mortality. Although several reports, usually representing the work of a few, uniquely qualified surgeons, have been published presenting negligible morbidity and mortality data, 4's some previous, community-based surveys have not documented comparable results. 6 Furthermore, previous publications have suggested that the results of carotid endarterectomy are superior when performed by surgeons on a regular basis and in hospitals in which this procedure represents a significant percentage of the operative volume. 7 To study the impact of these factors on the results of carotid endarterectomy in our own medical community, we retrospectively surveyed all hospitals serving the greater Cincinnati area for a continuous 12- month period and determined the perioperative morbidity and mortality. MATERIAL AND METHODS Study design. The medical record of every patient undergoing carotid endartercctomy between 911

2 912 Kempczinski, Brott, and Labutza ]ouma[ o( VASCULAR SURGERY Table I. Distribution of risk factors in symptomatic and asymptomatic patients Symptomatic Asvmptomatic Overall (%) (%) (%) Carotid bruit Hypertension 6 I Smoking Diabetes mellitis Myocardial infarct 21. i Angina pectoris Table II. Mean age and sexual distribution between symptomatic and asymptomatic patients Symptomatic Asymptomatic Age (yr) Male 61.1% 53.2% Female 38.9% 46.8% July 1, 1983 and June 30, 1984 in each of the 16 general hospitals serving the greater Cincinnati area were personally reviewed by one of us. All physician and nursing notes as well as anesthesia records, operative reports, and angiographic findings were examined whenever available. The pertinent information was abstracted and entered into a computerized data base for later review. Indications for operation. Patients with a fixed neurologic deficit, ipsilateral to the operated side, were considered to have "stroke" as the indication for that operation irrespective of the timing of the stroke relative to the procedure or the degree of residual deficit. Reversible ischemic neurologic deficits (RINDs) were classified as strokes. Similarly, hemispheric transient ischemic attacks or amaurosis fugax, ipsilateral to the operative side, were accepted as the indication for operation irrespective of how remotely they had occurred. However, if carotid endartercctomy were subsequently performed on the contralateral, asymptomatic side, that procedure was considered as being performed for an "asymptomatic" indication, despite the patient's previous contralateral symptoms. All carotid endartercctomies for ipsilateral stroke, hemispheric transient ischemic attack, amaurosis fugax, or RIND were included in the "symptomatic" group. All other indications, such as nonhemispheric, transient cerebral ischemia, dizziness alone, asymptomatic bruit or stenosis, or other unspecified indications, were considered "asymptomatic." A postoperative stroke was diagnosed when a new focal neurologic deficit, not present preopera- tively, was recorded in the hospital record on two consecutive days by either a physician or nurse. However, no focal deficits recorded by a nurse were included unless subsequently confirmed by a physician. Isolated facial numbness, hoarseness or tongue palsy, ipsilateral to the operated side, were attributed to cranial nerve injury and were not included in the computation of postoperative stroke. Surgeons. Sixty-one surgeons were represented in our study group. All surgeons were board-qualified and were classified into one of the following four categories: general surgeon, vascular surgeon, neurosurgeon, or cardiac surgeon. Since Certificates of Special Qualification in Vascular Surgery were not widely available at the time of our study, "vascular surgeons" were defined as physicians who had completed a recognized vascular fellowship or were currently members of either our regional vascular society (The Midwestern Vascular Surgical Society) or one of the national vascular societies (Society for Vascular Surgery or International Society for Cardiovascular Surgery). Board-qualified, practicing cardiothoracic surgeons were counted as "cardiac surgeons" even if they also met the criteria for "vascular surgeon." Neurosurgeons were physicians who were board-qualified in neurologic surgery and whose practice was limited to that specialty. Physicians who were boardqualified in general surgery and did not fit into one of the above three categories were classified as "general surgeons." Data analysis. Statistical analysis was performed by Rakesh Shulda, Ph.D., and Vicki Hertzberg, Ph.D. (Department of Environmental Health, University of Cincinnati Medical Center) with the chisquare test. RESULTS Altogether, 750 carotid endarterectomies were performed in 656 patients during the study period. The mean age of the patients was (SD) years with an almost equal sexual distribution (57% male, 43% female). Risk factors were equally distributed among symptomatic and asymptomatic pa-

3 Volume 3 Number 6 June 1986 Community-wide results of carotid endarterectomy TIA, 25-- hemisl)heric Strike 10 Amaurosis I ' RIND 0 ~ I Symptomatic n=369(49.2%) Bruit/ Stenosis Dizziness Asymptomatic n=381(50.8%) Fig. 1. Indications for carotid endarterectomy in symptomatic and asymptomatic patients. tients and were typical of an elderly patient population with disseminated atherosclerosis (Table I). Three hundred sixty-nine carotid endarterectomies were performed on "symptomatic" arteries, whereas 381 were performed for"asymptomatic" disease (Fig. 1). The age and sexual distribution among these two groups are listed in Table II. There was no statistically significant difference in the patient demographics or distribution of risk factors between the symptomatic and asymptomatic cohorts. Five hundred sixty-three patients had unilateral procedures. Ninety-three patients underwent bilateral, asynchronous carotid endarterectomy. Because one of these patients also underwent unilateral reoperation approximately 3 months later, 750 procedures were performed. The right carotid artery underwent repair in 363 operations and the left in 387 procedures. An intraoperative shunt was used during carotid occlusion in 33.5% of the operations. Although the combined morbidity and mortality rate was higher for shunted patients (9.6%) than for the nonshunted patients (6.2%), this difference did not achieve statistical significance. When a shunt was not used, mean occlusion time was 27.4 minutes. In 21 of the operations, an associated synchronous surgical procedure was performed: coronary artery bypass grafting in nine patients, vascular surgical reconstruction in eight, a general surgical procedure in two, and miscellaneous procedures in two other Overall, strokes occurred in 5.1% of all patients and 2.3% of all patients died. Symptomatic patients had a significantly (p < 0.05) higher risk of suffering postoperative stroke (6.5% vs. 3.7%), although the r % Vascular 1.9 ~.i/~ [] Death 2.6 [] Stroke ~ General Neuroaurgery Cardiac Fig. 2. Results of carotid endarterectomy on the basis of surgical specialty. risk of death was virtually identical (2.4% vs. 2.1%). Sixty-five percent (11 of 17) of the deaths that occurred were not related to postoperative stroke. If double counting of patients who suffered stroke and subsequently died was eliminated, stroke or death occurred in 7.8% of symptomatic patients and 5.3% of asymptomatic patients (no significance) or 6.5% of the total population. The mortality rate for those patients who sustained postoperative stroke was 15.8% and was virtually identical in symptomatic vs. asymptomatic patients. In 88% of the patients who suffered a perioperative stroke, the deficit was still present at the time of discharge from the hospital. In addition to the strokes, 22 patients (2.9%) had an episode of transient cerebral ischemia postoperatively. Therefore, the combined incidence of postoperative neurologic complications was 8% (60 of 750 procedures). When analyzed relative to the indication for operation, 36.8% of the patients who sustained a postoperative stroke had undergone operation for transient ischemic attacks, 26.3% for previous stroke, 21% for asymptomatic bruit or stenosis, 10.5% for dizziness, and 5.2% for nonspecific indications. The morbidity and mortality data relative to the specialty of the operating surgeon are summarized in Fig. 2. The number of surgeons in each category and their respective caseloads are listed in Table III.

4 914 Kempczinski, Brott, and Labu~ta Journal of" VASCULAR SURGERY Table III, Distribution of the cases by type of surgeon No. of operations No. of surgeons Average caseload % Asymptomatic on operated side Vascular General Neurosurgeon Cardiac No statistically significant difference in either postoperative stroke or death was noted among the four groups. The impact of each surgeon's operative caseload on the results after carotid endarterectomy are summarized in Fig. 3. Although combined morbidity appears to be lower in surgeons performing more than 50 carotid endarterectomies each year, this difference did not achieve statistical significance at the 0.05 level. Furthermore, surgeons in this group operated on more (61%) asymptomatic patients, who were at lower risk for postoperative stroke, compared with the two other groups in which only 44.1% of the patients were asymptomatic. Similarly, the 16 hospitals involved in the study were classified by the number of carotid endarterectomies performed in their operating rooms. Eleven hospitals averaged less than 50 carotid endartcrectomies each year. Four averaged between 50 and 100 procedures and a single hospital performed more than 100 procedures during the study period. There was no significant difference in either the postoperative stroke or death rate among the three groups of hospitals (Fig. 4). DISCUSSION Although individual, uniquely qualified surgeons s or specialized centers 4 have published low morbidity and mortality data occurring after carotid endartcrectomy, most such procedures are performed in community hospitals in which results may be more variable. In a previous publication, one of us (T.B.) first surveyed the results of carotid endartercctomy in the Cincinnati community during This represented the largest survey of its kind covering all carotid endarterectomies performed in a sophisticated, metropolitan medical community serving 1.4 million inhabitants. It demonstrated an overall perioperative stroke morbidity rate of 8.6% and a mortality rate of 2.8%. This report provided the impetus for our current study to examine the potential impact of surgical specialty and operative caseload on the results of surgery. It may be tempting to conclude that the high perioperative stroke rate we found was the result of an overly rigorous interpretation of "minor" ncurologic events. Unfortunately, 15.8% of the patients who suffered perioperative stroke died as a consequence of this complication and 88% of the survivors left the hospital with a residual deficit. Certainly, our meticulous review of every patient's hospital record could be expected to uncover a higher incidence of complications than might be expected from similar surveys made only on the basis of discharge face sheets. In fact, in our previous report, 59.4% ofperioperative strokes wcrc not recorded on the face sheet, s Consequently, other publications, on the basis of such documentation, may have significantly underestimated the true incidence of perioperativc stroke. We also do not believe that our results represent a unique, anomalously high morbidity and mortality. All surgeons in our study were Board-qualified. Furthermore, many of them have a longstanding interest in the treatment of cercbrovascular arterial occlusive disease and have made significant contributions to the medical literature on this subject. All participating hospitals were accredited by the Joint Committee on the Accreditation of Hospitals and had active quality assurance programs to define staff privileges for all surgical subspecialitics as well as morbidity and mortality conferences to monitor the results of surgical therapy. In a national review derived from hospital discharge summaries, Dyken and Pokras 2 reported a mortality rate of 2.8% among 82,000 patients undergoing carotid cndarterectomy during Slavish, Nicholas, and Gee 9 reviewed a recent community hospital experience with 743 carotid cndarterectomies and documented an operative mortality rate of 2.7%, whereas an earlier review of community hospitals found an operative mortality rate of 3.2% and a stroke rate of 10.7%. ~ West et al. t~ summarized the neurologic morbidity and mortality of 3233 carotid cndarterectomies in 16 published series (1968 to 1977) and found a 5.7% (range, 0.8% to 27%) perioperativc stroke rate and a 2.9% (range,

5 Volume 3 Number 6 June 1986 Community-wide results of carotid endarterectomy 915 % ~ 5-- 4r-- 3~ 2-- I Surgeons Avg. Cases n =<12 2.4! , N >50 Totah [] Death [] Stroke : N Fig. 3. Results of carotid endarterectomy grouped by each surgeon's annual caseload (less than 12 procedures/year, 12 to 50 procedures/year, and more than 50 procedures/ year). Number of surgeons in each group and their average caseload are also shown. 0% to 11%) procedure mortality. Given the reproducibility of our findings on two separate occasions more than 3 years apart and the concordance of those findings with the reports cited earlier, we believe that they are representative of those currently being achieved in large, metropolitan communities, such as our own, and probably the country as a whole. Although previous reports have suggested that the results of carotid endarterectomy are superior in the hands of surgeons who perform this operation as a regular part of their surgical practice, 7'H we could not document a relationship between outcome and either the formal specialty of the operating surgeon or his annual carotid endarterectomy caseload. Furthermore, because the 38 postoperative strokes in our study were equally distributed among 27 operating surgeons, no one surgeon was responsible for a disproportionate number of them. Certainly, the outstanding results reported by physicians with special interest in this type of surgery speak eloquently in support of the impact that experience and expertise can have on ensuring the most favorable outcome. However, within the limits of the present study, we could not document a clear relationship between volume and outcome. Other reports have also failed to demonstrate such a relationship. 9,~2 The question as to what constitutes an "adequate" caseload for carotid endarterectomy remains unan- % Avg. Cases Hospitals n ~-< ,( -, [] Death [] Stroke >100 Total Fig. 4. Results of carotid endarterectomy grouped by hospital's annual caseload (less than 50/year, 50 to 100/year, and more than 100/year). Number of hospitals and their average caseload in each category are also shown. swered. In fact, some critics might suggest that the average annual experience (21.2 cases each year) of even our busiest group (i.e., the vascular surgeons) might be inadequate to ensure the best results. Nevertheless, individual surgeons within each group had results comparable or superior to the best that have been reported, and the average caseloads we documented are probably more realistic estimates of the actual experience of most American surgeons practicing outside specialized centers. Although the incidence of perioperative stroke would appear to be largely under the control of the operating surgeon, the risk of postoperative death may not be. Nearly 65% of our deaths were not stroke-related and were invariably the resuk of cardiovascular complications that could not be directly attributed to any specific untoward perioperative event. Sixty-two percent of the carotid endarterectomies in our community were performed in five hospitals (average, 93 cases each year). Thus, most (69%) of our hospitals admitted patients for this procedure infrequently, averaging only 26 cases each year or roughly one case every 2 weeks. Nevertheless, the results of surgery in these institutions, when compared with hospitals in which this procedure was performed 3.5 times more often, were virtually identical (Fig. 4). This observation was at variance with some previous reports 7'H but was consistent with other regional surveys. I2

6 916 Kempczinski, Brott, and Labutta Jotlrna[ ol VASCULAR SURGERY More than 50% of the paticnts who undc~'ent carotid endarterectomy in our series were asymptomatic. The acceptable combined morbidity and mortality for carotid endarterectomy in such patients should be less than 3%. 13'14 Although the risk of perioperative stroke was significantly lower in our asymptomatic patients (3.7% vs. 6.5%), the risk of death remained virtually identical (2.4% vs. 2.1%) and the likelihood of such patients suffering either stroke or death remained nearly twice (5.3%) the recommended minimum. Since the natural histo U of asymptomatic carotid lesions remains unclear, thc combined morbidity that we documented warrants a careful reassessment of our managemcnt of thcsc patients. It also lends support to the need tbr a prospectivc, randomized, nationwide study comparing the efficacy of various forms of management for patients with asymptomatic carotid arterial disease. Important new work from the University of Washington has documented that only those asymptomatic patients with carotid stenosis in excess of 80% reduction of the diameter of the internal carotid artery appear to have a significantly increased risk of suffering neurologic complications from these lcsions.~5 These findings, if confirmed by other investigators, would suggest that all asymptomatic patients should be initially screened with a duplex scanner and that only those with high-grade stenosis be recommended for angiography and possible prophylactic endarterectomy. Because the difference between "acceptable" and "unacceptable" results after carotid endarterectomy in asymptomatic patients is so small and because it does not appear to be influenced by the operating surgeon's specialty or average annual caseload, it is imperative that each surgeon performing this procedure scrupulously monitor his own personal results and regularly hold them up against published benchmarks as a guide in his management of these challenging patients. We express our appreciation to all the Cincinnati surgeons who participated in this study. Their commitment to such a unique external review speaks eloquently in support of their determination to provide the highest qualiq, care to their patients. REFERENCES 1. United States Department of Health and Human Scrviccs. Detailed diagnosis and surgical procedures tbr patients discharged from short-stay hospitals. DHHS Publications (PHS) No. 82. US Government Printing Office, 1985: Dyken ML, Pokras R. The pertbrmance of endarterectomv for disease of the extracranial arteries of the head. Stroke 1984; 15: Dixon S, Pals SO, Raviola C, Gomes A, Machleder HI, Baker JD, Busuttil RW, Barker WF, Moore WS. Natural histoq' of nonstenotic, asymptomatic ulcerative lesions of the carotid artery. A further analysis. Arch Surg 1982; 117: Nicholls SC, Phillips DJ, Bergelin RO, Beach KW, Primozich JF, Strandness Jr DE. Carotid endarterectomy. Relationship of outcome to early restenosis. J VASC St:~G 1985; 2: Thompson JE. Carotid endarterectomy, the state of the art. Br J Surg 1983; 70: Easton JD, Sherman DG, Stroke and mortality rate in carotid endarterectomy: 228 consecutive operations. Stroke 1977; 8: Luft HS, Bunker JP, Enthoven AC. Should operations be regionalized? The empirical relation between surgical volume and mortality. N Engl J Med 1979; 301: Brott T, Thalinger K. The practice of carotid endarterectomy in a large metropolitan area. Stroke 1984; 15: Slavish LG, Nicholas GG, Gee W. Review of a communitw hospital experience with carotid endarterectomy. Stroke 1984; 15: Prioleau Jr WH, Aiken AF, Hairston P. Carotid endartcrectomy: Neurologic complications as related to surgical techniques. Ann Surg 1977; 185: West H, Burton R, Roon AJ, Malone JM, Goldstone J, Moore WS. Comparative risk of operation and expectant management tbr carotid artery disease. Stroke 1979; 10: Hertzer NR, Avellone JC, Farrell CJ, Plecha FR, Rhodes RS, Sharp WV, Wright GF. The risk of vascular surge D, in a metropolitan community. With observations on surgeon experience and hospital size. J VASC SURG 1984; 1: Jonas S, Hass WK. An approach to the maximal acceptablc stroke complication rate after surgery fbr transient cerebral ischemia (abstr). Stroke 1979; 10: Podore PC, DeWeese JA, May AG, Rob CG. Asymptomatic contralateral carotid artery stenosis: A five-year followup study fbllowing carotid endarterectomy. SurgeD~ 1980; 88: Roederer GO, Langlois YE, Lusiani L, Jager KA, Primozich JF, Lawrence RJ, Phillips DJ, Strandness Jr DE. Natural histo~ of carotid artery, disease on the side contralateral to endarterectomy. J VAse SuRe; 1984; 1:62-72.