- Email: [email protected]
Appendix J. Literature Review: Anesthesia
Appendix J. Literature Review: Anesthesia Purpose Anesthesia management may affect visual outcome and morbidity associated with cataract surgery in elderly patients. This literature search was undertaken to identify and review pertinent articles that could provide a solid basis for patient care recommendations relative to anesthesia management that would minimize morbidity and enhance visual outcome. The questions specified by the panel to be addressed in this review are detailed in the Results section.
Method A computerized literature search was performed by the National Library of Medicine to identify potentially pertinent articles published from January 1, 1975, through December 31, 1990, relating to anesthesia management and cataract surgery. (Chapter 2 and Appendix A give a detailed description of the search strategy.) This computer search identified 87 articles. Seven additional articles were identified from various sources, bringing the total number of identified articles to 94. There were 16 letters to the editor and 20 anesthesia case reports (1-5 cases), usually noting complications of retrobulbar block. Twenty-one articles described variations of techniques for local anesthesia, and 10 articles commented on a variety of medication management regimes. Seven articles that discussed general anesthesia for eye surgery were all published before 1978. Those that were published before 1975 were excluded. Of these 94 articles presented for preliminary review, only 10 were appropriate to be forwarded for more detailed content and methodologic review. The remaining 84 were rejected based on the inclusion criteria. Further review of the 10 revealed that 1 article did not meet the inclusion criteria (only 7 cases), and 1 article was found not to be relevant because it did not evaluate the topic of anesthesia. One other article (Whittpen, Rapoza, Sternberg et al., 1986), not identified in the initial search, was added because of its possible relevance. Therefore, nine Prepared by John V. Donlon, MD; Denis M. O'Day, MD, FACS, Panel Chair: HenryS. Sacks, MD, PhD; and Dinah Reitman, MPS, who comprised the literature review group.
articles were included and underwent detailed review. Attachment J-2 at the end of this appendix summarizes the article selection process. The papers were reviewed by John V. Donlon, MD, Kenneth Zahl, MD, and Gerald Wolf, MD, according to the specific questions relating anesthesia care to visual outcome and morbidity and using a content evaluation form (Attachment J-1). The methodology review and evidence tables were constructed by HenryS. Sacks, MD, PhD, and Dinah Reitman, MPS, based on Woolf (1990). The information abstracted for the tables included the following: • Study identification. • Study design. • Population. • Exposure/intervention. • Outcome measures. • Findings. • Comments.
Summary of Methodology Review The following is a summary by article of the methodology review of the literature on anesthesia for cataract surgery. (See also Evidence Table J-1 at the end of this appendix.)
Redmond and Dallas (1990) This was a retrospective study of 100 patients who underwent extracapsular cataract extraction (ECCE) surgery and lens implantation under local anesthesia. A total of 50 patients had retrobulbar injection, the established technique, and 50 patients had subconjunctival infiltration (nonretrobulbar). This study was not designed as a prospective randomized controlled trial, which was a major flaw. The authors claimed that there were no significant differences in mean age or sex ratio between the groups, although the numbers of males vs. females were not given. No discussion of differences in baseline visual acuity, systemic disease, etc., were provided. The authors claimed no differences in mean improvement of visual acuity or astigmatism between the two groups after surgery, although no statistical analysis was presented, and there was no discussion of a possible type 2 error. An unknown number of patients were excluded from the analysis, including patients with retrobulbar hemorrhages that were not enumerated as 2015
Ophthalmology
Volume 100, Supplement, August 1993
complications and patients in whom vision did not improve "owing to previously unrecognized macular or optic nerve pathology." Excluding these patients after the results of the study were known was obviously a subject of bias and a major flaw. The complications that were enumerated were too small in number to be significant.
Smith (1990) This was an uncontrolled trial (termed by the author a feasibility study) of nonretrobulbar local anesthesia for ECCE performed in 175 eyes of 165 patients. The investigator provided some comparison of postoperative visual acuity between the study group and 10 patients who had general anesthesia and 3 other patients who had retrobulbar anesthesia. The author was cautious enough to specify that this comparison was purely illustrative, with no obvious differences in the outcomes between the two groups, and made no attempt at statistical significance. There was no discussion of baseline visual acuity, nor any comparison of acuity before and after surgery, which could have included confidence intervals of the differences. The number of complications were few enough to warrant a full-scale randomized controlled trial. It was not clear why it took 165 patients to decide on the feasibility of this technique before such a trial could be undertaken.
Whitsett, Balyeat, and McClure (1990) This was a double-blind randomized controlled trial of peribulbar vs. retrobulbar anesthesia for ECCE in 100 patients. There is no baseline information about these patients and no information about inclusion/exclusion criteria of the study nor about where the investigation took place. The evaluation was based mostly on the efficacy of the anesthetic block, for which no statistically significant differences were found, although there was no discussion of statistical methods. It was noted that no serious or potentially life-threatening complications were observed, but there was no discussion of preoperative or postoperative visual acuity, visual function, astigmatism, or less serious side effects. There was a cautious note that the sample size was small and that a larger trial was indicated. Moreover, an acknowledgment was made of the possibility of bias influencing the results because one of the observers involved in grading the effectiveness of the anesthesia had not been blinded. There were, however, multiple observers in the grading process, for which this study deserves credit.
2025
Hamilton, Gimbel, and Strunin (1988) This study is a consecutive series of 12,000 patients who had surgery during the period January 1984-April 1988. Five different types of blocks were used for anesthesia for cataract surgery. The block types were chronologically assigned A through E. Block A was a retrobulbar plus 7th nerve block (Nadbath or Atkinson technique); block B was a retrobulbar with no 7th nerve blocking (which the authors called an "increased volume block"); block C was a "dual peribulbar block" with two intraorbital injections; block D was a dual peribulbar block with one injection intraorbitally and one through the upper eyelid; block E was what the authors called a "custom" block, that is, a peribulbar block plus a retrobulbar block if necessary. The outcome measures used were: (1) block score (i.e., how well the block performed in terms of anesthesia) and (2) complications from the block. The authors found that blocks A and B achieved the best scores but that they also had the highest proportion of complications. Patients receiving blocks A and B had brain stem anesthesia and other central nervous system spread. In addition, "breakthrough pain" was significantly greater with blocks A and B compared with C, D, and E, and block A was significantly worse than block B. "Muscle paresis" (not defined by the authors) was highest with block B, and shivering was highest with block E. Other factors could possibly have produced these results, however. The majority of complications were seen with blocks A and B, and these blocks were administered chronologically early in the experience of the surgeons. Thus, there is no way to differentiate between a chronological effect (such as a learning curve) and an effect of the anesthesia itself. The Nadbath technique, used in an undisclosed group of patients in block A, may have been responsible for the brain stem effect in this group. In addition, the study was not masked, so classification of complications was potentially subject to bias. The number of complications is small in each group, making statistical comparisons between groups difficult. Moreover, the complications were not precisely defined. Nonetheless, it is a reasonable hypothesis that an increased rate of complications, particularly involving the central nervous system, might be associated with the techniques used in blocks A and B, as reported in this study. There were no data on visual outcome or the relationship between visual outcome and complications of the surgery.
Cataract Management Guideline Panel • Management of Functional Impairment Due to Cataract in Adults
Karhunen and Orko (1981) This was a randomized control trial of 150 female elderly patients undergoing cataract extraction. One hundred patients had local anesthesia, and 50 had general anesthesia. Fifty patients receiving local anesthesia and 50 receiving general were given pethidine and atropine. The remaining 50 receiving local anesthesia were given these same drugs plus droperidol. Differences in postoperative nausea and vomiting were charted over the ensuing 28 hours, although the four time periods reported in the article seemed to be determined retrospectively, a method that could have biased the results. Baseline ages, weights, and drug dosages were provided, but there was no discussion of the differences, although they appeared to be similar across groups. Although common in all groups, there was generally less nausea and vomiting in the droperidol group than in the other two groups for the first two time periods; statistical significance (determined by X2 analysis) was reached only between the droperidol and the general anesthesia groups. However, the p values were not given for the nonsignificant differences, and the possibility of a type 2 error was not discussed. Additionally, supplemental analgesia, sedation, and antiemetics were given to several patients. There was a statistically significant difference in that less analgesia and sedation were given to the droperidol group than to the local anesthesia group without droperidol. This may have influenced the results of the study. Additionally, some patients in each group were given antiemetics (including droperidol) postoperatively, which may have also affected the outcome. In addition to a multivariate analysis, which may have controlled for some of these confounding variables, there should have been a fourth group in the original design (general anesthesia with droperidol).
Weiss and Deichman (1989) This was a prospective randomized controlled trial of 79 consecutive patients who received cataract extractions with intraocular lens implants: 40 patients who received retrobulbar injections, and 39 who received periocular injections. There was no discussion of the patient population at baseline. Although more patients in the retrobulbar group required supplemental anesthesia, this difference was not statistically significant (p>O.IO). Only one patient in the retrobulbar group had any complication (tachycardia). Visual acuity, visual function, and astigmatism were not considered in evaluating the differences in these anesthesia techniques. The mean amount of chemosis was less (p<0.001) in the
retrobulbar group than in the periocular. The authors concluded that periocular anesthesia was an effective and perhaps a safer alternative to retrobulbar. There was no discussion of the possibility of a type 2 error, although the authors did advocate the need for further large-scale studies.
Backer, Tinker, Robertson et al. (1980) From this retrospective series of more than 10,000 ophthalmologic procedures managed with local and/or retrobulbar block, 195 patients who underwent 288 operations and who had previously suffered a myocardial infarction were identified. There were no postoperative reinfarctions or deaths in this group. During the same period, 21 patients with prior myocardial infarction had 26 operations under general anesthesia. There were no postoperative reinfarctions in this smaller group either. Another group of 255 patients with documented coronary artery disease manifested by angina, but with no previous myocardial infarction, underwent 373 operations under local anesthesia; there were two postoperative myocardial infarctions. Since these numbers were too small to allow any conclusions to be drawn, the authors compared their results with those from a prior study of 587 nonophthalmic operations using general or major regional anesthesia on patients with prior myocardial infarction. Their results yielded a reinfarction rate of 6.1 percent when including abdominal, intrathoracic, and other operations lasting more than 3 hours. When these patients were eliminated from the analysis, the reinfarction rate was 3.6 percent, which the authors claimed was still statistically significant in X 2 analysis. The authors concluded, therefore, that local anesthesia and/or retrobulbar block for ophthalmic surgery did not pose significant risk for reinfarction. This study lacked prospective validation. The retrospective comparison between ophthalmic and nonophthalmic surgery and between local and general anesthesia in two different studies involved too many variables to draw any meaningful conclusions. Although a prospective randomized controlled trial to answer this question might require many thousands of patients, perhaps a meta-analysis of numerous studies of local vs. general anesthesia for ophthalmic studies with this endpoint in mind could provide more useful data.
Meyers (1979) This was a retrospective evaluation of a clinical series of 1,000 patients who underwent intraocular surgery under local anesthesia without a standby anesthesiologist vs. another series of 300 patients 2035
Ophthalmology
Volume 100, Supplement, August 1993
with standby. The incidence of complications was 5.6 percent and 13 percent in the no-standby and standby groups, respectively, although the seriousness of the complications in the no-standby group was greater. To determine the cost effectiveness of having a standby anesthesiologist present during local anesthesia would require a prospectively designed randomized controlled trial with an accompanying cost analysis. Comparing retrospective results from two clinical series is subject to numerous biases and confounding variables, even though the author noted similarities in the patients' ages as well as operating room conditions and personnel. For example, no data were given on preexisting conditions of these two populations, which may have been different. It is interesting that more numerous, though less serious, complications occurred in the standby group, but the author was quick to conclude that the standby anesthesiologist enabled these complications to be quickly controlled, probably a biased assessment in favor of standby.
Whittpen, Rapoza, Sternberg et al. (1986) This is a consecutive series of patients receiving retrobulbar injections administered by a number of different surgeons. The 3,123 patients were separated into two groups. Group 1 received 2-percent lidocaine, bupivacaine, and hyaluronidase, and group 2 received 4-percent lidocaine and the same doses of bupivacaine and hyaluronidase as group 1. The group 1 patients had a variety of ophthalmic procedures, including cataract surgery, whereas the group 2 patients all underwent retinal reattachment or vitrectomy. The major outcome measure of interest was respiratory arrest. Two of the 2,235 group 1 patients (2-percent lidocaine) and 7 of the 888 group 2 patients (4-percent lidocaine) had respiratory arrest. Four of the seven group 2 patients had diabetes mellitus. No other significant differences were seen between the two groups except for the difference in the lidocaine concentration. Serum levels of lidocaine and bupivacaine in the respiratory arrest patients were compared with levels in 20 "control" patients, that is, patients who did not have respiratory arrest. No differences in serum levels between patients with respiratory arrest and the control patients were seen. These results, overall, indicate that the patients receiving the higher concentration of lidocaine are at increased risk for respiratory arrest. Unfortunately, respiratory arrest was not precisely defined in this study, nor was the study specific to cataract patients. The patients in groups 1 and 2 had different underlying reasons for surgery and may also
2045
have had different comorbidities. These comorbidities may be associated with risk of respiratory arrest. In addition, the injections were carried out by different surgeons. Therefore, it is not possible to differentiate between the effect of the technique of injecting the anesthesia and the effect of the anesthesia itself. In terms of the serum values measured in the "patients" and "controls," it was not clear who comprised the 20 controls and whether they came from the same original group of patients.
Results Question 1: Is there evidence to support or refute an association between visual outcome of cataract surgery and the choice of general vs. local anesthesia? No evidence was found in the literature addressing this question. Question 2: Is there evidence to support or refute an association between visual outcome of cataract surgery and the choice of local anesthesia, retrobulbar vs. peribulbar? No evidence was found in the literature addressing this question. One article (Redmond and Dallas, 1990) suggested that visual acuity (Snellen) was unaffected by general, retrobulbar block, or subconjunctival injection. However, as noted above, this study had a number of serious methodologic flaws. Question 3: Is there evidence to support or refute an association between visual outcome of cataract surgery and the identity and credentials of the individual monitoring the patient? Question 4: Is there evidence to support or refute an association between visual outcome of cataract surgery and the identity and credentials of the individual administering retrobulbar injection? (Consider anesthesiologist, nurse anesthetist, ophthalmologist, technician, other.) No evidence was found in the literature addressing these questions. Question 5: Is there evidence to support or refute an association between anesthesia morbidity and choice of general vs. local anesthesia? One paper (Karhunen and Orko, 1981) found that nausea and vomiting may be less likely following local anesthesia. As noted above, this paper had significant flaws. Question 6: Is there evidence to support or refute an association between anesthesia morbidity and choice of local anesthesia, retrobulbar vs. peribulbar?
Cataract Management Guideline Panel • Management of Functional Impairment Due to Cataract in Adults Although three articles (Hamilton, Gimbel, and Strunin, 1988; Weiss and Deichman, 1989; Whitsett, Balyeat, and McClure, 1990) seemed to indicate that a peribulbar technique can be as effective as retrobulbar but results in fewer complications and the hypothesis is reasonable, none of the data was statistically significant. Another study (Whittpen, Rapoza, Sternberg et al., 1986) suggested that patients receiving 4-percent lidocaine by retrobulbar injection were at increased risk of respiratory arrest compared with patients receiving a 2-percent preparation. The study was marred by general methodologic deficiencies that made the conclusions difficult to interpret, though it is intuitively appealing.
Question 7: Is there evidence to support or refute an association between anesthesia morbidity and the identity and credentials of the individual monitoring the patient? (Consider anesthesiologist, nurse anesthetist, operating ophthalmologist, other.) One study (Meyers, 1979) suggested that there was an increased incidence of severe cardiovascular complications in unmonitored patients. There were a number of methodologic flaws in this study, which was a retrospective evaluation of a clinical series. The conclusions were uncertain and difficult to interpret. Question 8: Is there evidence to support or refute an association between anesthesia morbidity and the identity and credentials of the individual administering retrobulbar or peribulbar injection? (Consider anesthesiologist, nurse anesthetist, ophthalmologist, technician, other.) No articles specifically compared morbidity and complications based on the credentials of the person performing the local anesthesia. Question 9: Is there information on cost and/or cost-benefit issues? No articles were found addressing cost or costbenefit issues.
Discussion Of the 94 articles identified by the National Library of Medicine computer search and other sources, only 9 met the search criteria and were further considered by content and methodologic reviewers for detailed analysis. Results of the review of these articles revealed that only three were randomized controlled trials. The remainder were either uncontrolled trials or clinical series. There were no well-designed prospective studies that compared general anesthesia vs. local anesthesia with
anesthesia morbidity. No studies associated morbidity with the identity of the person monitoring or administering the local anesthesia injection. Two articles mentioned cost effectiveness in passing, but no study specifically analyzed in detail the costs and benefits of general vs. local anesthesia. Only two articles (Redmond and Dallas, 1990; Smith, 1990) mentioned visual outcome. There was no information to associate visual outcome with choice of general or local anesthesia. Most articles noted complications of surgery or anesthesia but were of poor quality.
Conclusions The literature yielded little or no valid data for clear, specific patient care recommendations. There is an urgent need for large, well-designed, prospective studies that adequately address these questions or the impact of anesthesia choices on visual outcome and morbidity in the elderly patient having cataract surgery.
Reference In addition to the articles on the literature review lists for anesthesia, the following reference is cited in this appendix. Woolf SH. Interim manual on clinical practice guideline development. Rockville (MD): Agency for Health Care Policy and Research; 1990.
Literature Review Lists Literature Included This list comprises the nine articles that were found to be relevant and were further reviewed for content and methodologic rigor. Backer CL, Tinker JH, Robertson DM, Vliestra RE. Myocardial reinfarction following local anesthesia for ophthalmic surgery. Anesth Analg 1980 Apr;59(4):257-62. Hamilton RC, Gimbel HV, Strunin L. Regional anaesthesia for 12,000 cataract extraction and intraocular lens implantation procedures. Can J Anaesth 1988 Nov;35(6):615-23. Karhunen U, Orko R. Nausea and vomiting after local anesthesia for cataract extraction in elderly female patients-effect of droperidol premedication. Ophthalmic Surg 1981 Nov;12(11):810-2. Meyers EF. Problems during eye surgery. Under local anesthesia, is standby necessary? Anesth Rev 1979 Jul;6(7):23-5. Redmond RM, Dallas NL. Extracapsular cataract extraction under local anaesthesia without retrobulbar
2055
Ophthalmology
Volume 100, Supplement, August 1993
injection. Br J Ophthalmol 1990 Apr;74(4):203-4. Comment. In: Br J Ophthalmol 1990 Oct;74(10):639. Smith R. Cataract extraction without retrobulbar anaesthetic injection. Br J Ophthalmol 1990 Apr;74(4):205-7. Weiss JL, Deichman CB. A comparison of retrobulbar and periocular anesthesia for cataract surgery. Arch Ophthalmol 1989 Jan;107(1):96-8. Whitsett JC, Balyeat HD, McClure B. Comparison of oneinjection-site peribulbar anesthesia and retrobulbar anesthesia. J Cataract Refract Surg 1990 Mar;16(2):243-5. Comment. In: J Cataract Refract Surg 1990 Jul;16(4):5278. Whittpen JR, Rapoza P, Sternberg P, Kuwashima L, Saklad J, Arnall P. Respiratory arrest following retrobulbar anesthesia. Ophthalmology 1986 Jul;93(7):867-70.
2065
Literature Excluded This list comprises the two articles excluded because they did not meet the inclusion criteria. Ingram RM, Banerjee D, Traynar MJ, Thompson RK. Day-case cataract surgery. Trans Ophthalmol Soc UK 1980 Apr;lOO(Pt 1):205-9. Schneider ME, Milstein DE, Oyakawa RT, Ober RR, Campo R. Ocular perforation from a retrobulbar injection. Am J Ophthalmol 1988 Jul 15;106(1):35-40.
Cataract Management Guideline Panel • Management of Functional Impairment Due to Cataract in Adults
Evidence Table J-1. Anesthesia Reference: Karhunen and Orko, 1981. Study Design: Randomized controlled trial. Study Population: 150 female patients, age= 72±7. Exposure/Intervention: Atropine (local), pethidine with/without droperidol (local), atropine (general), pethidine (general). Outcome Measures: Nausea and vomiting. Findings: Droperidol appeared to reduce nausea in early postop time periods, but only statistically significant between local group with droperidol and general group without. Comments: Although this was a randomized trial, there were too many confounding variables (should have been 2x2 factorial design; general vs. local with and without droperidol). No discussion of type 2 error as possible reason for lack of differences in some analyses. Needs multivariate analysis because of possible effects of other analgesics given. Reference: Redmond and Dallas, 1990. Study Design: Clinical series--comparative study. Study Population: 100 Caucasian patients, mean age= 76.6 (57 to 96 range), age and sex ratios equivalent. Exposure/Intervention: Retrobulbar vs. nonretrobulbar (subconjunctival bupivacaine) anesthesia. Outcome Measures: Visual acuity; astigmatism complications. Findings: No statistically significant differences in improvement of visual acuity or postop astigmatism. Complications included vitreous loss (3 retro, 2 nonretro), iris prolapse (1 retro), 2 wound leaks (retro), 2 residual cortex (nonretro). Surgeon's opinion-retro caused more stress than nonretro. Comments: Not a prospective randomized study. An unknown number of cases were excluded from analysis retrospectively. This could have been subject to bias. No discussion of type 2 error. No discussion of baseline differences in visual acuity or systemic disease. Reference: Smith, 1990. Study Design: Uncontrolled trial (pseudocontrol groups). Study Population: 208 eyes (193 patients, 118 female, 75 male, mean age 70.5, range 46 to 90, SO 14.5). Main series, 165 patients (175 eyes); 26 patients (30 eyes). Exposure/Intervention: Nonretrobulbar anesthesia. Pseudocontrols: 30 eyes (general anesthesia); 3 eyes (retrobulbar). Outcome Measures: Visual acuity complications. Findings: There were few complications with nonretrobulbar anesthesia: 1 had convulsive hand movement during surgery; vitreous loss in 3 eyes; 1 had pain. No vitreous loss in general anesthesia cases. No obvious differences in postop visual acuity in 3 groups of patients, but no statistical analysis was performed. Nonretrobulbar technique was considered acceptable. Comments: Not designed as a controlled trial, but used two comparison groups as a reference. This was really a phase 1/11 study and needs confirmation by a larger randomized controlled trial. Reference: Whitsett, Balyeat, and McClure, 1990. Study Design: Double-blind randomized controlled trial. Study Population: 100 patients (no description). Exposure/Intervention: Retrobulbar vs. peribulbar local anesthesia. Outcome Measures: Anesthesia efficacy (globe akinesia, lid akinesia, globe anesthesia). Supplemental anesthesia. Complications. Findings: No statistically significant differences between groups for any outcomes measured. "No serious or potentially catastrophic complications." Comments: Although this was a prospective double-blind randomized controlled trial, there was no discussion of patient populations, pre/postop visual acuity, function, or astigmatism. No discussion of statistical methods used. Authors did acknowledge small sample size and possible bias because of unblinded observers. Reference: Hamilton, Gimbel, and Strunin, 1988. Study Design: Consecutive series 1/2/84 to 4/8/88. Block type chronologically assigned A through E. Study Population: 12,000 patients; about 70% were ~70 years old; female, male rates, 62:38. Exposure/Intervention: Blocks: A) retrobulbar plus 7th nerve block (N=3,595); B) increased volume retrobulbar block (N=1 ,640); C) dual peribulbar block (2 intraorbital) (N=3,478); D) dual peribulbar block (1 intraorbital, 1 through upper eyelid) (N=2,226); E) custom block (peribulbar plus retro is necessary) (N=1 ,061 ). Outcome Measures: 1) Block score. 2) Complications: • brain stem anesthesia; • other CNS spread; • spread to contralateral orbit; • moderate retrobulbar hemorrhage; • eyelid conjunctivitis and peribulbar ecchymoses; • extraocular muscle paresis; • shivering; • scleral perforation; • optic atrophy; • vasovagal problems; • "breakthrough" pain. Findings: 1) Blocks A and B achieved best scores. 2) 6/3,595 block A and 2/1 ,640 block B patients had brain stem anesthesia; 2 and 3 patients, respectively, had other CNS spread. 3) 15/1 ,640 block B patients had extraocular muscle paresis (compared with 4, 3, and 1 with blocks C, D, and E). 4) Shivering highest with block E (0.64%). 5) Vasovagal problems ranged from 0.5 to 0.85%. 6) "Breakthrough" pain significantly greater with A and B compared with C, D, and E; A was significantly worse than B. Comments: • Study not masked so classification of complications subject to bias. • Numbers of complications small in each group; statistical comparisons between groups difficult. • No data on visual outcome or relationship between complications and postop vision. • Complications in chronologically "early" block methods: no way to differentiate between chronologie effect (learning cause) and anesthetic effect.
2075
Ophthalmology
Volume 100, Supplement, August 1993
Reference: Backer, Tinker, Robertson et al., 1980. Study Design: Case series (comparative). Study Population: 1a) 288 ophthalmic operations (local anesthesia) with prior myocardial infarction (195 patients); 69% males; 90% >60 years of age; mean age 72 years; 18% had diabetes; 40% had angina. 1b) 26 additional operations (21 patients) had general anesthesia (prior myocardial infarction). 2) Other series of 587 nonophthalmic operations; 79% males; mean age=67 years; 21% had diabetes; 34% had angina. Exposure/Intervention: 1a) Ophthalmic surgery under local anesthesia. 1b) Ophthalmic surgery under general anesthesia. 2) Nonophthalmic surgery under general or major regional anesthesia. Outcome Measures: Complications (postop reinfarction). Findings: • No postop reinfarctions among local/retrobulbar or general anesthesia ophthalmic operations. • 6.1% reinfarction rate in nonophthalmic surgery group. • 3.3% reinfarction rate when only procedures Q hr are included; 3.6% when "risky" upper abdominal, thoracic, or greater vessel operations of ~3 hr are eliminated. Comments: All retrospective comparisons. Too few ophthalmic procedures under general anesthesia with prior myocardial infarctions to expect any reinfarctions. Comparison of ophthalmology surgery with other types is subject to many confounding variables. This question may be researched through a meta-analysis. Will need many prospective clinical trials to find enough reinfarctions to draw conclusions. Reference: Meyers, 1979. Study Design: Case series (comparative). Study Population: Approximately 1,300 patients; mean age=68 years. Exposure/Intervention: Intraocular surgery-300 patients with standby anesthesiologist, 1,000 patients without standby (both groups had retrobulbar block). Outcome Measures: Complications. Findings: Without standby: 5.6% complication rate (mostly life threatening); 13% complications in nonstandby group, but excluded patients with tachycardia who were taking anticholinergic medications. Authors claim standby group did better. Comments: Needs validation through randomized controlled trial. No discussion of baseline differences other than similarities in age. Why were tachycardia patients excluded from analysis of standby group? No indication that standby group was superior. Reference: Whittpen, Rapoza, Sternberg et al., 1986. Study Design: Consecutive series; different surgeons. Study Population: 3,123 patients requiring retrobulbar injections. Group 1 patients (N=2,235) had a variety of procedures including cataract surgery. All group 2 patients (N=388) underwent retinal reattachment or vitrectomy. Serum values of 20 "control" patients (noncases) also used. Exposure/Intervention: Group 1: 2% lidocaine+ 0.75% bupivacaine +hyaluronidase. Group 2: 4% lidocaine+ 0.75% bupivacaine + hyaluronidase. All patients received preop sedation. Outcome Measures: Respiratory arrest. Findings: 1) 2/2,235 group 1 patients had respiratory arrest (p=0.003) vs. 7/888 group 2 patients. 2) 4/7 group 2 patients had diabetes mellitus. 3) No sign differences in age, sex between groups. 4) No increase in serum levels of lidocaine or bupivacaine in patients compared with 20 "control" patients. Comments: • Patients in groups 1 and 2 had different surgeries and may have had different comorbidities associated with risk of respiratory arrest. • Not clear if 20 "controls" were noncases from groups 1 and 2. • Not specific to cataract patients. • Respiratory arrest not defined. • Retrobulbar injection not defined or carried out by same surgeons. Cannot differentiate between technique effect and anesthesia effect. Reference: Weiss and Deichman, 1989. Study Design: Single-blind randomized controlled trial (surgeon blinded). Study Population: 79 patients, unknown characteristics. Exposure/Intervention: Retrobulbar vs. periocular anesthesia. Outcome Measures: Complications: supplemental anesthesia, chemosis. Findings: Only 1 patient in retrobulbar group had complication (tachycardia). 8/39 vs. 11/39 in retrobulbar vs. periocular patients required supplemental anesthesia (p>0.1 0, not significant); less chemosis in retrobulbar group (p<0.001 ); otherwise, no statistically significant difference between groups. Comments: No discussion of population characteristics or differences thereof. No consideration of type 2 error. No evaluation of pre/postop visual acuity, visual function, astigmatism. Note: CNS = Central nervous system. SD = Standard deviation.
2085
Cataract Management Guideline Panel • Management of Functional Impairment Due to Cataract in Adults
Attachment J-1. Content Evaluation Form: Anesthesia Article ID #: Reviewer ID #: Review Date: • Nature of the study population. Check ALL that apply: _ _ _ Population-based sample ___ Statistical sample of cataract patients _ _ _ Other, describe_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ • Does the study contain information to support or refute an association between visual outcome of cataract surgery and 1. Choice of general vs. local anesthesia? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
2. Choice of local anesthesia: retrobulbar vs. peribulbar vs. other
(specify, other) _ _ _ _ _ _ _ __ _ Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
3. The identity and credentials of the individual monitoring the patient? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
4. The identity and credentials of the individual administering retrobulbar injection (i.e., anesthesiologist, nurse anesthetist, ophthalmologist, technician, other)? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
• Does the study contain information to support or refute an association between anesthesia morbidity and
1. Choice of general vs. local anesthesia?
_Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
2. Choice of local anesthesia, retrobulbar vs. peribulbar vs. other (specify, other) _ _ _ _ _ _ _ __ _ Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
How many different study groups were there and what types of injections were given? No Yes Retrobulbar Yes No Peribulbar No Yes Subconjunctival Other _ _ _ __ Yes No Were any adverse events observed more frequently in certain study groups? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ _ __ 2095
Ophthalmology
Volume 100, Supplement, August 1993
3. The identity and credentials of the individual monitoring the patient? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __ 4. The identity and credentials of the individual administering retrobulbar injection (i.e., anesthesiologist, nurse anesthetist, ophthalmologist, technician, other)? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ _ __ • Does the study contain information on cost or cost/benefit issues? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ _ __ • If article characteristics are not covered in this questionnaire, please indicate by checking here: _ __
2105
Cataract Management Guideline Panel • Management of Functional Impairment Due to Cataract in Adults
Attachment J-2. Article Selection Process: Anesthesia National Library of Medicine Search:
87 potentially pertinent articles relating to anesthesia
J.
J, J, J, J,
7 articles added : Identified from various sources
f- f - f- f- f -
PLUS
f - f-
f-
94 potentially pertinent articles on anesthesia
J.
J, J, J, J,
84
articles rejected: Did not meet inclusion criteria for anesthesia
---7 ---7 ---7 ---7 ---7
L ESS
---7 ---7
---7
10 potentially pertinent articles on anesthesia
J, J, J, J, J,
2 articles excluded : Did not meet inclusion criteria
---7 ---7 ---7 ---7 ---7
L ESS
---7 ---7
---7
8 articles included and reviewed for content and methodologic quality
J,
1 article added : Identified as pertinent
J.
J, J, J,
f - f - f- f - f -
PLUS
f - f-
f-
9 articles
New total for included articles
2115
Method A computerized literature search was performed by the National Library of Medicine to identify potentially pertinent articles published from January 1, 1975, through December 31, 1990, relating to anesthesia management and cataract surgery. (Chapter 2 and Appendix A give a detailed description of the search strategy.) This computer search identified 87 articles. Seven additional articles were identified from various sources, bringing the total number of identified articles to 94. There were 16 letters to the editor and 20 anesthesia case reports (1-5 cases), usually noting complications of retrobulbar block. Twenty-one articles described variations of techniques for local anesthesia, and 10 articles commented on a variety of medication management regimes. Seven articles that discussed general anesthesia for eye surgery were all published before 1978. Those that were published before 1975 were excluded. Of these 94 articles presented for preliminary review, only 10 were appropriate to be forwarded for more detailed content and methodologic review. The remaining 84 were rejected based on the inclusion criteria. Further review of the 10 revealed that 1 article did not meet the inclusion criteria (only 7 cases), and 1 article was found not to be relevant because it did not evaluate the topic of anesthesia. One other article (Whittpen, Rapoza, Sternberg et al., 1986), not identified in the initial search, was added because of its possible relevance. Therefore, nine Prepared by John V. Donlon, MD; Denis M. O'Day, MD, FACS, Panel Chair: HenryS. Sacks, MD, PhD; and Dinah Reitman, MPS, who comprised the literature review group.
articles were included and underwent detailed review. Attachment J-2 at the end of this appendix summarizes the article selection process. The papers were reviewed by John V. Donlon, MD, Kenneth Zahl, MD, and Gerald Wolf, MD, according to the specific questions relating anesthesia care to visual outcome and morbidity and using a content evaluation form (Attachment J-1). The methodology review and evidence tables were constructed by HenryS. Sacks, MD, PhD, and Dinah Reitman, MPS, based on Woolf (1990). The information abstracted for the tables included the following: • Study identification. • Study design. • Population. • Exposure/intervention. • Outcome measures. • Findings. • Comments.
Summary of Methodology Review The following is a summary by article of the methodology review of the literature on anesthesia for cataract surgery. (See also Evidence Table J-1 at the end of this appendix.)
Redmond and Dallas (1990) This was a retrospective study of 100 patients who underwent extracapsular cataract extraction (ECCE) surgery and lens implantation under local anesthesia. A total of 50 patients had retrobulbar injection, the established technique, and 50 patients had subconjunctival infiltration (nonretrobulbar). This study was not designed as a prospective randomized controlled trial, which was a major flaw. The authors claimed that there were no significant differences in mean age or sex ratio between the groups, although the numbers of males vs. females were not given. No discussion of differences in baseline visual acuity, systemic disease, etc., were provided. The authors claimed no differences in mean improvement of visual acuity or astigmatism between the two groups after surgery, although no statistical analysis was presented, and there was no discussion of a possible type 2 error. An unknown number of patients were excluded from the analysis, including patients with retrobulbar hemorrhages that were not enumerated as 2015
Ophthalmology
Volume 100, Supplement, August 1993
complications and patients in whom vision did not improve "owing to previously unrecognized macular or optic nerve pathology." Excluding these patients after the results of the study were known was obviously a subject of bias and a major flaw. The complications that were enumerated were too small in number to be significant.
Smith (1990) This was an uncontrolled trial (termed by the author a feasibility study) of nonretrobulbar local anesthesia for ECCE performed in 175 eyes of 165 patients. The investigator provided some comparison of postoperative visual acuity between the study group and 10 patients who had general anesthesia and 3 other patients who had retrobulbar anesthesia. The author was cautious enough to specify that this comparison was purely illustrative, with no obvious differences in the outcomes between the two groups, and made no attempt at statistical significance. There was no discussion of baseline visual acuity, nor any comparison of acuity before and after surgery, which could have included confidence intervals of the differences. The number of complications were few enough to warrant a full-scale randomized controlled trial. It was not clear why it took 165 patients to decide on the feasibility of this technique before such a trial could be undertaken.
Whitsett, Balyeat, and McClure (1990) This was a double-blind randomized controlled trial of peribulbar vs. retrobulbar anesthesia for ECCE in 100 patients. There is no baseline information about these patients and no information about inclusion/exclusion criteria of the study nor about where the investigation took place. The evaluation was based mostly on the efficacy of the anesthetic block, for which no statistically significant differences were found, although there was no discussion of statistical methods. It was noted that no serious or potentially life-threatening complications were observed, but there was no discussion of preoperative or postoperative visual acuity, visual function, astigmatism, or less serious side effects. There was a cautious note that the sample size was small and that a larger trial was indicated. Moreover, an acknowledgment was made of the possibility of bias influencing the results because one of the observers involved in grading the effectiveness of the anesthesia had not been blinded. There were, however, multiple observers in the grading process, for which this study deserves credit.
2025
Hamilton, Gimbel, and Strunin (1988) This study is a consecutive series of 12,000 patients who had surgery during the period January 1984-April 1988. Five different types of blocks were used for anesthesia for cataract surgery. The block types were chronologically assigned A through E. Block A was a retrobulbar plus 7th nerve block (Nadbath or Atkinson technique); block B was a retrobulbar with no 7th nerve blocking (which the authors called an "increased volume block"); block C was a "dual peribulbar block" with two intraorbital injections; block D was a dual peribulbar block with one injection intraorbitally and one through the upper eyelid; block E was what the authors called a "custom" block, that is, a peribulbar block plus a retrobulbar block if necessary. The outcome measures used were: (1) block score (i.e., how well the block performed in terms of anesthesia) and (2) complications from the block. The authors found that blocks A and B achieved the best scores but that they also had the highest proportion of complications. Patients receiving blocks A and B had brain stem anesthesia and other central nervous system spread. In addition, "breakthrough pain" was significantly greater with blocks A and B compared with C, D, and E, and block A was significantly worse than block B. "Muscle paresis" (not defined by the authors) was highest with block B, and shivering was highest with block E. Other factors could possibly have produced these results, however. The majority of complications were seen with blocks A and B, and these blocks were administered chronologically early in the experience of the surgeons. Thus, there is no way to differentiate between a chronological effect (such as a learning curve) and an effect of the anesthesia itself. The Nadbath technique, used in an undisclosed group of patients in block A, may have been responsible for the brain stem effect in this group. In addition, the study was not masked, so classification of complications was potentially subject to bias. The number of complications is small in each group, making statistical comparisons between groups difficult. Moreover, the complications were not precisely defined. Nonetheless, it is a reasonable hypothesis that an increased rate of complications, particularly involving the central nervous system, might be associated with the techniques used in blocks A and B, as reported in this study. There were no data on visual outcome or the relationship between visual outcome and complications of the surgery.
Cataract Management Guideline Panel • Management of Functional Impairment Due to Cataract in Adults
Karhunen and Orko (1981) This was a randomized control trial of 150 female elderly patients undergoing cataract extraction. One hundred patients had local anesthesia, and 50 had general anesthesia. Fifty patients receiving local anesthesia and 50 receiving general were given pethidine and atropine. The remaining 50 receiving local anesthesia were given these same drugs plus droperidol. Differences in postoperative nausea and vomiting were charted over the ensuing 28 hours, although the four time periods reported in the article seemed to be determined retrospectively, a method that could have biased the results. Baseline ages, weights, and drug dosages were provided, but there was no discussion of the differences, although they appeared to be similar across groups. Although common in all groups, there was generally less nausea and vomiting in the droperidol group than in the other two groups for the first two time periods; statistical significance (determined by X2 analysis) was reached only between the droperidol and the general anesthesia groups. However, the p values were not given for the nonsignificant differences, and the possibility of a type 2 error was not discussed. Additionally, supplemental analgesia, sedation, and antiemetics were given to several patients. There was a statistically significant difference in that less analgesia and sedation were given to the droperidol group than to the local anesthesia group without droperidol. This may have influenced the results of the study. Additionally, some patients in each group were given antiemetics (including droperidol) postoperatively, which may have also affected the outcome. In addition to a multivariate analysis, which may have controlled for some of these confounding variables, there should have been a fourth group in the original design (general anesthesia with droperidol).
Weiss and Deichman (1989) This was a prospective randomized controlled trial of 79 consecutive patients who received cataract extractions with intraocular lens implants: 40 patients who received retrobulbar injections, and 39 who received periocular injections. There was no discussion of the patient population at baseline. Although more patients in the retrobulbar group required supplemental anesthesia, this difference was not statistically significant (p>O.IO). Only one patient in the retrobulbar group had any complication (tachycardia). Visual acuity, visual function, and astigmatism were not considered in evaluating the differences in these anesthesia techniques. The mean amount of chemosis was less (p<0.001) in the
retrobulbar group than in the periocular. The authors concluded that periocular anesthesia was an effective and perhaps a safer alternative to retrobulbar. There was no discussion of the possibility of a type 2 error, although the authors did advocate the need for further large-scale studies.
Backer, Tinker, Robertson et al. (1980) From this retrospective series of more than 10,000 ophthalmologic procedures managed with local and/or retrobulbar block, 195 patients who underwent 288 operations and who had previously suffered a myocardial infarction were identified. There were no postoperative reinfarctions or deaths in this group. During the same period, 21 patients with prior myocardial infarction had 26 operations under general anesthesia. There were no postoperative reinfarctions in this smaller group either. Another group of 255 patients with documented coronary artery disease manifested by angina, but with no previous myocardial infarction, underwent 373 operations under local anesthesia; there were two postoperative myocardial infarctions. Since these numbers were too small to allow any conclusions to be drawn, the authors compared their results with those from a prior study of 587 nonophthalmic operations using general or major regional anesthesia on patients with prior myocardial infarction. Their results yielded a reinfarction rate of 6.1 percent when including abdominal, intrathoracic, and other operations lasting more than 3 hours. When these patients were eliminated from the analysis, the reinfarction rate was 3.6 percent, which the authors claimed was still statistically significant in X 2 analysis. The authors concluded, therefore, that local anesthesia and/or retrobulbar block for ophthalmic surgery did not pose significant risk for reinfarction. This study lacked prospective validation. The retrospective comparison between ophthalmic and nonophthalmic surgery and between local and general anesthesia in two different studies involved too many variables to draw any meaningful conclusions. Although a prospective randomized controlled trial to answer this question might require many thousands of patients, perhaps a meta-analysis of numerous studies of local vs. general anesthesia for ophthalmic studies with this endpoint in mind could provide more useful data.
Meyers (1979) This was a retrospective evaluation of a clinical series of 1,000 patients who underwent intraocular surgery under local anesthesia without a standby anesthesiologist vs. another series of 300 patients 2035
Ophthalmology
Volume 100, Supplement, August 1993
with standby. The incidence of complications was 5.6 percent and 13 percent in the no-standby and standby groups, respectively, although the seriousness of the complications in the no-standby group was greater. To determine the cost effectiveness of having a standby anesthesiologist present during local anesthesia would require a prospectively designed randomized controlled trial with an accompanying cost analysis. Comparing retrospective results from two clinical series is subject to numerous biases and confounding variables, even though the author noted similarities in the patients' ages as well as operating room conditions and personnel. For example, no data were given on preexisting conditions of these two populations, which may have been different. It is interesting that more numerous, though less serious, complications occurred in the standby group, but the author was quick to conclude that the standby anesthesiologist enabled these complications to be quickly controlled, probably a biased assessment in favor of standby.
Whittpen, Rapoza, Sternberg et al. (1986) This is a consecutive series of patients receiving retrobulbar injections administered by a number of different surgeons. The 3,123 patients were separated into two groups. Group 1 received 2-percent lidocaine, bupivacaine, and hyaluronidase, and group 2 received 4-percent lidocaine and the same doses of bupivacaine and hyaluronidase as group 1. The group 1 patients had a variety of ophthalmic procedures, including cataract surgery, whereas the group 2 patients all underwent retinal reattachment or vitrectomy. The major outcome measure of interest was respiratory arrest. Two of the 2,235 group 1 patients (2-percent lidocaine) and 7 of the 888 group 2 patients (4-percent lidocaine) had respiratory arrest. Four of the seven group 2 patients had diabetes mellitus. No other significant differences were seen between the two groups except for the difference in the lidocaine concentration. Serum levels of lidocaine and bupivacaine in the respiratory arrest patients were compared with levels in 20 "control" patients, that is, patients who did not have respiratory arrest. No differences in serum levels between patients with respiratory arrest and the control patients were seen. These results, overall, indicate that the patients receiving the higher concentration of lidocaine are at increased risk for respiratory arrest. Unfortunately, respiratory arrest was not precisely defined in this study, nor was the study specific to cataract patients. The patients in groups 1 and 2 had different underlying reasons for surgery and may also
2045
have had different comorbidities. These comorbidities may be associated with risk of respiratory arrest. In addition, the injections were carried out by different surgeons. Therefore, it is not possible to differentiate between the effect of the technique of injecting the anesthesia and the effect of the anesthesia itself. In terms of the serum values measured in the "patients" and "controls," it was not clear who comprised the 20 controls and whether they came from the same original group of patients.
Results Question 1: Is there evidence to support or refute an association between visual outcome of cataract surgery and the choice of general vs. local anesthesia? No evidence was found in the literature addressing this question. Question 2: Is there evidence to support or refute an association between visual outcome of cataract surgery and the choice of local anesthesia, retrobulbar vs. peribulbar? No evidence was found in the literature addressing this question. One article (Redmond and Dallas, 1990) suggested that visual acuity (Snellen) was unaffected by general, retrobulbar block, or subconjunctival injection. However, as noted above, this study had a number of serious methodologic flaws. Question 3: Is there evidence to support or refute an association between visual outcome of cataract surgery and the identity and credentials of the individual monitoring the patient? Question 4: Is there evidence to support or refute an association between visual outcome of cataract surgery and the identity and credentials of the individual administering retrobulbar injection? (Consider anesthesiologist, nurse anesthetist, ophthalmologist, technician, other.) No evidence was found in the literature addressing these questions. Question 5: Is there evidence to support or refute an association between anesthesia morbidity and choice of general vs. local anesthesia? One paper (Karhunen and Orko, 1981) found that nausea and vomiting may be less likely following local anesthesia. As noted above, this paper had significant flaws. Question 6: Is there evidence to support or refute an association between anesthesia morbidity and choice of local anesthesia, retrobulbar vs. peribulbar?
Cataract Management Guideline Panel • Management of Functional Impairment Due to Cataract in Adults Although three articles (Hamilton, Gimbel, and Strunin, 1988; Weiss and Deichman, 1989; Whitsett, Balyeat, and McClure, 1990) seemed to indicate that a peribulbar technique can be as effective as retrobulbar but results in fewer complications and the hypothesis is reasonable, none of the data was statistically significant. Another study (Whittpen, Rapoza, Sternberg et al., 1986) suggested that patients receiving 4-percent lidocaine by retrobulbar injection were at increased risk of respiratory arrest compared with patients receiving a 2-percent preparation. The study was marred by general methodologic deficiencies that made the conclusions difficult to interpret, though it is intuitively appealing.
Question 7: Is there evidence to support or refute an association between anesthesia morbidity and the identity and credentials of the individual monitoring the patient? (Consider anesthesiologist, nurse anesthetist, operating ophthalmologist, other.) One study (Meyers, 1979) suggested that there was an increased incidence of severe cardiovascular complications in unmonitored patients. There were a number of methodologic flaws in this study, which was a retrospective evaluation of a clinical series. The conclusions were uncertain and difficult to interpret. Question 8: Is there evidence to support or refute an association between anesthesia morbidity and the identity and credentials of the individual administering retrobulbar or peribulbar injection? (Consider anesthesiologist, nurse anesthetist, ophthalmologist, technician, other.) No articles specifically compared morbidity and complications based on the credentials of the person performing the local anesthesia. Question 9: Is there information on cost and/or cost-benefit issues? No articles were found addressing cost or costbenefit issues.
Discussion Of the 94 articles identified by the National Library of Medicine computer search and other sources, only 9 met the search criteria and were further considered by content and methodologic reviewers for detailed analysis. Results of the review of these articles revealed that only three were randomized controlled trials. The remainder were either uncontrolled trials or clinical series. There were no well-designed prospective studies that compared general anesthesia vs. local anesthesia with
anesthesia morbidity. No studies associated morbidity with the identity of the person monitoring or administering the local anesthesia injection. Two articles mentioned cost effectiveness in passing, but no study specifically analyzed in detail the costs and benefits of general vs. local anesthesia. Only two articles (Redmond and Dallas, 1990; Smith, 1990) mentioned visual outcome. There was no information to associate visual outcome with choice of general or local anesthesia. Most articles noted complications of surgery or anesthesia but were of poor quality.
Conclusions The literature yielded little or no valid data for clear, specific patient care recommendations. There is an urgent need for large, well-designed, prospective studies that adequately address these questions or the impact of anesthesia choices on visual outcome and morbidity in the elderly patient having cataract surgery.
Reference In addition to the articles on the literature review lists for anesthesia, the following reference is cited in this appendix. Woolf SH. Interim manual on clinical practice guideline development. Rockville (MD): Agency for Health Care Policy and Research; 1990.
Literature Review Lists Literature Included This list comprises the nine articles that were found to be relevant and were further reviewed for content and methodologic rigor. Backer CL, Tinker JH, Robertson DM, Vliestra RE. Myocardial reinfarction following local anesthesia for ophthalmic surgery. Anesth Analg 1980 Apr;59(4):257-62. Hamilton RC, Gimbel HV, Strunin L. Regional anaesthesia for 12,000 cataract extraction and intraocular lens implantation procedures. Can J Anaesth 1988 Nov;35(6):615-23. Karhunen U, Orko R. Nausea and vomiting after local anesthesia for cataract extraction in elderly female patients-effect of droperidol premedication. Ophthalmic Surg 1981 Nov;12(11):810-2. Meyers EF. Problems during eye surgery. Under local anesthesia, is standby necessary? Anesth Rev 1979 Jul;6(7):23-5. Redmond RM, Dallas NL. Extracapsular cataract extraction under local anaesthesia without retrobulbar
2055
Ophthalmology
Volume 100, Supplement, August 1993
injection. Br J Ophthalmol 1990 Apr;74(4):203-4. Comment. In: Br J Ophthalmol 1990 Oct;74(10):639. Smith R. Cataract extraction without retrobulbar anaesthetic injection. Br J Ophthalmol 1990 Apr;74(4):205-7. Weiss JL, Deichman CB. A comparison of retrobulbar and periocular anesthesia for cataract surgery. Arch Ophthalmol 1989 Jan;107(1):96-8. Whitsett JC, Balyeat HD, McClure B. Comparison of oneinjection-site peribulbar anesthesia and retrobulbar anesthesia. J Cataract Refract Surg 1990 Mar;16(2):243-5. Comment. In: J Cataract Refract Surg 1990 Jul;16(4):5278. Whittpen JR, Rapoza P, Sternberg P, Kuwashima L, Saklad J, Arnall P. Respiratory arrest following retrobulbar anesthesia. Ophthalmology 1986 Jul;93(7):867-70.
2065
Literature Excluded This list comprises the two articles excluded because they did not meet the inclusion criteria. Ingram RM, Banerjee D, Traynar MJ, Thompson RK. Day-case cataract surgery. Trans Ophthalmol Soc UK 1980 Apr;lOO(Pt 1):205-9. Schneider ME, Milstein DE, Oyakawa RT, Ober RR, Campo R. Ocular perforation from a retrobulbar injection. Am J Ophthalmol 1988 Jul 15;106(1):35-40.
Cataract Management Guideline Panel • Management of Functional Impairment Due to Cataract in Adults
Evidence Table J-1. Anesthesia Reference: Karhunen and Orko, 1981. Study Design: Randomized controlled trial. Study Population: 150 female patients, age= 72±7. Exposure/Intervention: Atropine (local), pethidine with/without droperidol (local), atropine (general), pethidine (general). Outcome Measures: Nausea and vomiting. Findings: Droperidol appeared to reduce nausea in early postop time periods, but only statistically significant between local group with droperidol and general group without. Comments: Although this was a randomized trial, there were too many confounding variables (should have been 2x2 factorial design; general vs. local with and without droperidol). No discussion of type 2 error as possible reason for lack of differences in some analyses. Needs multivariate analysis because of possible effects of other analgesics given. Reference: Redmond and Dallas, 1990. Study Design: Clinical series--comparative study. Study Population: 100 Caucasian patients, mean age= 76.6 (57 to 96 range), age and sex ratios equivalent. Exposure/Intervention: Retrobulbar vs. nonretrobulbar (subconjunctival bupivacaine) anesthesia. Outcome Measures: Visual acuity; astigmatism complications. Findings: No statistically significant differences in improvement of visual acuity or postop astigmatism. Complications included vitreous loss (3 retro, 2 nonretro), iris prolapse (1 retro), 2 wound leaks (retro), 2 residual cortex (nonretro). Surgeon's opinion-retro caused more stress than nonretro. Comments: Not a prospective randomized study. An unknown number of cases were excluded from analysis retrospectively. This could have been subject to bias. No discussion of type 2 error. No discussion of baseline differences in visual acuity or systemic disease. Reference: Smith, 1990. Study Design: Uncontrolled trial (pseudocontrol groups). Study Population: 208 eyes (193 patients, 118 female, 75 male, mean age 70.5, range 46 to 90, SO 14.5). Main series, 165 patients (175 eyes); 26 patients (30 eyes). Exposure/Intervention: Nonretrobulbar anesthesia. Pseudocontrols: 30 eyes (general anesthesia); 3 eyes (retrobulbar). Outcome Measures: Visual acuity complications. Findings: There were few complications with nonretrobulbar anesthesia: 1 had convulsive hand movement during surgery; vitreous loss in 3 eyes; 1 had pain. No vitreous loss in general anesthesia cases. No obvious differences in postop visual acuity in 3 groups of patients, but no statistical analysis was performed. Nonretrobulbar technique was considered acceptable. Comments: Not designed as a controlled trial, but used two comparison groups as a reference. This was really a phase 1/11 study and needs confirmation by a larger randomized controlled trial. Reference: Whitsett, Balyeat, and McClure, 1990. Study Design: Double-blind randomized controlled trial. Study Population: 100 patients (no description). Exposure/Intervention: Retrobulbar vs. peribulbar local anesthesia. Outcome Measures: Anesthesia efficacy (globe akinesia, lid akinesia, globe anesthesia). Supplemental anesthesia. Complications. Findings: No statistically significant differences between groups for any outcomes measured. "No serious or potentially catastrophic complications." Comments: Although this was a prospective double-blind randomized controlled trial, there was no discussion of patient populations, pre/postop visual acuity, function, or astigmatism. No discussion of statistical methods used. Authors did acknowledge small sample size and possible bias because of unblinded observers. Reference: Hamilton, Gimbel, and Strunin, 1988. Study Design: Consecutive series 1/2/84 to 4/8/88. Block type chronologically assigned A through E. Study Population: 12,000 patients; about 70% were ~70 years old; female, male rates, 62:38. Exposure/Intervention: Blocks: A) retrobulbar plus 7th nerve block (N=3,595); B) increased volume retrobulbar block (N=1 ,640); C) dual peribulbar block (2 intraorbital) (N=3,478); D) dual peribulbar block (1 intraorbital, 1 through upper eyelid) (N=2,226); E) custom block (peribulbar plus retro is necessary) (N=1 ,061 ). Outcome Measures: 1) Block score. 2) Complications: • brain stem anesthesia; • other CNS spread; • spread to contralateral orbit; • moderate retrobulbar hemorrhage; • eyelid conjunctivitis and peribulbar ecchymoses; • extraocular muscle paresis; • shivering; • scleral perforation; • optic atrophy; • vasovagal problems; • "breakthrough" pain. Findings: 1) Blocks A and B achieved best scores. 2) 6/3,595 block A and 2/1 ,640 block B patients had brain stem anesthesia; 2 and 3 patients, respectively, had other CNS spread. 3) 15/1 ,640 block B patients had extraocular muscle paresis (compared with 4, 3, and 1 with blocks C, D, and E). 4) Shivering highest with block E (0.64%). 5) Vasovagal problems ranged from 0.5 to 0.85%. 6) "Breakthrough" pain significantly greater with A and B compared with C, D, and E; A was significantly worse than B. Comments: • Study not masked so classification of complications subject to bias. • Numbers of complications small in each group; statistical comparisons between groups difficult. • No data on visual outcome or relationship between complications and postop vision. • Complications in chronologically "early" block methods: no way to differentiate between chronologie effect (learning cause) and anesthetic effect.
2075
Ophthalmology
Volume 100, Supplement, August 1993
Reference: Backer, Tinker, Robertson et al., 1980. Study Design: Case series (comparative). Study Population: 1a) 288 ophthalmic operations (local anesthesia) with prior myocardial infarction (195 patients); 69% males; 90% >60 years of age; mean age 72 years; 18% had diabetes; 40% had angina. 1b) 26 additional operations (21 patients) had general anesthesia (prior myocardial infarction). 2) Other series of 587 nonophthalmic operations; 79% males; mean age=67 years; 21% had diabetes; 34% had angina. Exposure/Intervention: 1a) Ophthalmic surgery under local anesthesia. 1b) Ophthalmic surgery under general anesthesia. 2) Nonophthalmic surgery under general or major regional anesthesia. Outcome Measures: Complications (postop reinfarction). Findings: • No postop reinfarctions among local/retrobulbar or general anesthesia ophthalmic operations. • 6.1% reinfarction rate in nonophthalmic surgery group. • 3.3% reinfarction rate when only procedures Q hr are included; 3.6% when "risky" upper abdominal, thoracic, or greater vessel operations of ~3 hr are eliminated. Comments: All retrospective comparisons. Too few ophthalmic procedures under general anesthesia with prior myocardial infarctions to expect any reinfarctions. Comparison of ophthalmology surgery with other types is subject to many confounding variables. This question may be researched through a meta-analysis. Will need many prospective clinical trials to find enough reinfarctions to draw conclusions. Reference: Meyers, 1979. Study Design: Case series (comparative). Study Population: Approximately 1,300 patients; mean age=68 years. Exposure/Intervention: Intraocular surgery-300 patients with standby anesthesiologist, 1,000 patients without standby (both groups had retrobulbar block). Outcome Measures: Complications. Findings: Without standby: 5.6% complication rate (mostly life threatening); 13% complications in nonstandby group, but excluded patients with tachycardia who were taking anticholinergic medications. Authors claim standby group did better. Comments: Needs validation through randomized controlled trial. No discussion of baseline differences other than similarities in age. Why were tachycardia patients excluded from analysis of standby group? No indication that standby group was superior. Reference: Whittpen, Rapoza, Sternberg et al., 1986. Study Design: Consecutive series; different surgeons. Study Population: 3,123 patients requiring retrobulbar injections. Group 1 patients (N=2,235) had a variety of procedures including cataract surgery. All group 2 patients (N=388) underwent retinal reattachment or vitrectomy. Serum values of 20 "control" patients (noncases) also used. Exposure/Intervention: Group 1: 2% lidocaine+ 0.75% bupivacaine +hyaluronidase. Group 2: 4% lidocaine+ 0.75% bupivacaine + hyaluronidase. All patients received preop sedation. Outcome Measures: Respiratory arrest. Findings: 1) 2/2,235 group 1 patients had respiratory arrest (p=0.003) vs. 7/888 group 2 patients. 2) 4/7 group 2 patients had diabetes mellitus. 3) No sign differences in age, sex between groups. 4) No increase in serum levels of lidocaine or bupivacaine in patients compared with 20 "control" patients. Comments: • Patients in groups 1 and 2 had different surgeries and may have had different comorbidities associated with risk of respiratory arrest. • Not clear if 20 "controls" were noncases from groups 1 and 2. • Not specific to cataract patients. • Respiratory arrest not defined. • Retrobulbar injection not defined or carried out by same surgeons. Cannot differentiate between technique effect and anesthesia effect. Reference: Weiss and Deichman, 1989. Study Design: Single-blind randomized controlled trial (surgeon blinded). Study Population: 79 patients, unknown characteristics. Exposure/Intervention: Retrobulbar vs. periocular anesthesia. Outcome Measures: Complications: supplemental anesthesia, chemosis. Findings: Only 1 patient in retrobulbar group had complication (tachycardia). 8/39 vs. 11/39 in retrobulbar vs. periocular patients required supplemental anesthesia (p>0.1 0, not significant); less chemosis in retrobulbar group (p<0.001 ); otherwise, no statistically significant difference between groups. Comments: No discussion of population characteristics or differences thereof. No consideration of type 2 error. No evaluation of pre/postop visual acuity, visual function, astigmatism. Note: CNS = Central nervous system. SD = Standard deviation.
2085
Cataract Management Guideline Panel • Management of Functional Impairment Due to Cataract in Adults
Attachment J-1. Content Evaluation Form: Anesthesia Article ID #: Reviewer ID #: Review Date: • Nature of the study population. Check ALL that apply: _ _ _ Population-based sample ___ Statistical sample of cataract patients _ _ _ Other, describe_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ • Does the study contain information to support or refute an association between visual outcome of cataract surgery and 1. Choice of general vs. local anesthesia? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
2. Choice of local anesthesia: retrobulbar vs. peribulbar vs. other
(specify, other) _ _ _ _ _ _ _ __ _ Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
3. The identity and credentials of the individual monitoring the patient? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
4. The identity and credentials of the individual administering retrobulbar injection (i.e., anesthesiologist, nurse anesthetist, ophthalmologist, technician, other)? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
• Does the study contain information to support or refute an association between anesthesia morbidity and
1. Choice of general vs. local anesthesia?
_Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
2. Choice of local anesthesia, retrobulbar vs. peribulbar vs. other (specify, other) _ _ _ _ _ _ _ __ _ Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __
How many different study groups were there and what types of injections were given? No Yes Retrobulbar Yes No Peribulbar No Yes Subconjunctival Other _ _ _ __ Yes No Were any adverse events observed more frequently in certain study groups? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ _ __ 2095
Ophthalmology
Volume 100, Supplement, August 1993
3. The identity and credentials of the individual monitoring the patient? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ __ 4. The identity and credentials of the individual administering retrobulbar injection (i.e., anesthesiologist, nurse anesthetist, ophthalmologist, technician, other)? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ _ __ • Does the study contain information on cost or cost/benefit issues? _Yes _No If Yes, describe _ _ _ _ _ _ _ _ _ _ _ _ __ • If article characteristics are not covered in this questionnaire, please indicate by checking here: _ __
2105
Cataract Management Guideline Panel • Management of Functional Impairment Due to Cataract in Adults
Attachment J-2. Article Selection Process: Anesthesia National Library of Medicine Search:
87 potentially pertinent articles relating to anesthesia
J.
J, J, J, J,
7 articles added : Identified from various sources
f- f - f- f- f -
PLUS
f - f-
f-
94 potentially pertinent articles on anesthesia
J.
J, J, J, J,
84
articles rejected: Did not meet inclusion criteria for anesthesia
---7 ---7 ---7 ---7 ---7
L ESS
---7 ---7
---7
10 potentially pertinent articles on anesthesia
J, J, J, J, J,
2 articles excluded : Did not meet inclusion criteria
---7 ---7 ---7 ---7 ---7
L ESS
---7 ---7
---7
8 articles included and reviewed for content and methodologic quality
J,
1 article added : Identified as pertinent
J.
J, J, J,
f - f - f- f - f -
PLUS
f - f-
f-
9 articles
New total for included articles
2115