National Outcomes of Cataract Extraction

N ationa! Outcomes of Cataract Extraction

Increased Risk of Retinal Complications Associated with N d: Y AG Laser Capsulotomy

Jonathan C. Javitt, MD, MPH,1.2.3 James M. Tielsch, PhD, 2 Joseph K. Canner, MHS,3 Margaret M. Kolb, DrPH,3·4 Alfred Sommer, MD, MHS,5 Earl P. Steinberg, MD, MPP,3.4.6 on behalf of the Cataract Patient Outcomes Research Team

Purpose: The authors studied 57,103 randomly selected Medicare beneficiaries who underwent extracapsular cataract extraction in 1986 or 1987 to determine the possible association between performance of neodymium (Nd):YAG laser capsulotorny and the risk of subsequent retinal break or detachment.

Methods: Cases of cataract surgery were identified from Medicare claims submitted in 1986 and 1987 and were followed through the end of 1988. Episodes of cataract surgery, posterior capsulotomy, and retinal complications were ascertained based on procedure and diagnosis codes listed in physician bills and hospital discharge records. Lifetable and Cox's proportional hazards models were used to analyze the risk of retinal detachment or break in patients undergoing and not undergoing capsulotomy during the period of observation.

Results: Of the 57,103 persons identified as having undergone extracapsular cataract extraction in 1986 or 1987, 13,709 subsequently underwent Nd:YAG laser capsulotomy between 1986 and 1988. A total of 337 persons had aphakic or pseudophakic retinal detachments between 1986 and 1988 and an additional 194 underwent repair of a retinal break. Proportional hazards modeling shows a 3.9-fold increase in the risk of retinal break or detachment among those who underwent capsulotomy (95% confidence interval: 2.89 to 5.25). Younger patient age, male sex, and white race also were associated with increased risk of retinal complications after extracapsular cataract extraction.

Conclusion: The authors conclude that there is a statistically significant increase in the risk of retinal detachment or break in those patients who undergo caps ulotomy after cataract extraction. Therefore, capsulotomy should be deferred until the patient's impairment caused by capsular opacification warrants the increased risk of retinal complications associated with performance of capsulotomy. Ophthalmology 1992;99: 1487-1498

Originally received: February 12, 1992. Revision accepted: June 4, 1992.

3 The Johns Hopkins Program for Medical Technology and Practice Assessment, Baltimore.

I Worthen Center for Eye Care Research, Center for Sight, Georgetown University Medical Center, Washington, DC.

2 Dana Center for Preventive Ophthalmology, Wilmer Ophthalmological Institute, Johns Hopkins University, Baltimore.

4 Department of Health Policy and Management, The Johns Hopkins School of Public Health, Baltimore.

5 Office of the Dean, The Johns Hopkins School of Public Health, Baltimore.

1487

Ophthalmology Volume 99, Number 10, October 1992

Posterior capsular opacity has been reported to occur postoperatively in 18% to 50% of patients who undergo extracapsular cataract extraction.1.2 Capsulotomy using the neodymium:Y AG (Nd:YAG) laser was originally described by Aron-Rosa and co-workers3 and has become the predominant approach for treating visually significant posterior capsular opacity. Although the procedure is generally safe and effective, Steinert and co-workers4 have recently reported that 0.89% of patients experience retinal detachment in the first year after capsulotomy. They also summarized an extensive literature search documenting that retinal detachment may occur in 0.08% to 3.6% of patients who undergo capsulotomy.5-18 The reports are based either on case series of retinal detachments or retrospective surveys of individuals who have undergone capsulotomy. The periods of observation and completeness of follow-up in these studies vary considerably.

We undertook an analysis of Medicare claims data to examine the association between performance ofNd:YAG capsulotomy and the risk of subsequent aphakic or pseudophakic retinal detachment. Our analysis is based on a large, representative sample of Medicare beneficiaries who underwent extracapsular cataract extraction (including phacoemulsification) in 1986 or 1987.

Methods

The Medicare statistical system maintained by the Health Care Financing Administration (HCFA) contains computerized information on Medicare beneficiaries, including data on demographic characteristics, hospital discharges, outpatient hospital procedures, physician services, skilled nursing care, and home health services. 19 To identifya representative sample of extracapsular cataract surgery, we searched the claims records of a random 5% sample of Medicare beneficiaries (approximately 1.5 million persons) for hospital records and physicians' bills related to cataract surgery performed during 1986 or 1987.

Identification of Cases of Cataract Surgery, Capsulotomy, and Retinal Complications

We identified individuals who possibly underwent cataract surgery (cases) on the basis of an International Classification of Diseases (ICD-9) code2o or HCFA Common Procedural Classification System (HCPCS) code21 indicative of cataract surgery (a complete list of codes used in

6 Department of Medicine, The Johns Hopkins School of Medicine, Baltimore.

Presented at the American Academy of Ophthalmology Annual Meeting, Anaheim, October 1991.

Supported in part by POI HS06280 from the Agency for Health Care Policy and Research, Rockville, Maryland, RO I EY08805 from the National Eye Institute/NIH, Bethesda, Maryland, and SIO RR06758 from the division of Research Resources/NIH, Bethesda, Maryland.

Reprint requests to Jonathan C. Javitt, MD, MPH, Worthen Center for Eye Care Research, Center for Sight, 3800 Reservoir Rd, NW, Washington, DC 20007.

1488

the identification of possible cataract cases and complications is available on request). HCPCS codes are classified as levels 1, 2, and 3, with HCPCS level 1 being identical to the Common Procedural Terminology (CPT_4)22 maintained by the American Medical Association. HCPCS level 2 codes are those designated by HCFA for nationwide use and were provided to us by the HCFA Bureau of Program Operations.

The specific HCFA data files searched were the Medicare Provider Annual Review (MEDPAR) file, the Part B Medicare Annual Data (BMAD) Beneficiary File, and the Part B Hospital Outpatient Facility file. The BMAD file contains physician bills and facility bills from ambulatory surgical centers, while the MEDPAR and hospital outpatient files contain inpatient and outpatient hospital (not ambulatory surgical centers) facility bills, respectively.

Until 1990, Medicare carriers were allowed to create their own HCPCS level 3 procedure codes, referred to as "local codes," instead of using HCPCS level 1 and 2 codes in Part B claims (i.e., bills submitted by a physician or ambulatory surgical center). To identify all possible episodes of cataract surgery, capsulotomy, and retinal detachment in the BMAD file, we examined all records from the 5% random sample in which the provider was identified as an ophthalmologist. This search yielded approximately 500,000 records with procedures designated by local codes. In most cases, the procedure was identified manually, based on lists of local codes submitted by Medicare carriers to HCF A. There were 35 local codes used for ophthalmologic procedures during 1986 and 1987 that could not be identified in this manner. The carriers using those codes were contacted by telephone and in all cases were able to provide definitions. While this procedure yielded no additional cases of cataract surgery, 8 codes, representing 345 cases of Nd:Y AG capsulotomy, were identified.

Using the above approach, we identified 80,299 possible cases of cataract surgery (Fig 1). We excluded 1558 persons who did not have continuous Part A and Part B

Excl . USIODS

Total Possible Cases 80,299 Nol boIh pari Aa B eIlP>le; N-1558

HMO Nembon; N-1616 Not u.s. Rcoidom; N-554

+ I!aIIded '-"-of diubiIity "" !!SRD; N-1511 Nol 66 or older; N-3061

Total Eligible Cases I 72,111

IntnocopouJar atracticm; N-849S

+ Typo of IIIIIJCIY 1IIIIpOCiIicd; N-2223 CoDcumDl 0IIIeri0t ~tm:tomy; N-554

Possible Cases of CombiIIod procedure; N-l664

&ttacapsular Cawact &traction

59,175

I

.' Levd 3 c.s; N=2072 I Levelland 2 Cases

57,103

Figure 1. Flowchart of steps in assembly of cataract cohort for data analysis. Because individuals may be excluded for multiple reasons, total number of exclusions is larger than total persons excluded.

Javitt et al . National Outcom,;Cs of Cataract Extraction

Table 1. Classification of Possible Cataract Surgical Episodes

Level I (Most Stringent) Must Meet Either Criterion 1 or 2: 1. Primary Evidence: A paid surgeon's bill for cataract surgery plus one or more of the following

pieces of confirmatory data: A. A paid bill or discharge record from an ambulatory surgery center, hospital outpatient

department or short stay hospital indicating cataract extraction and/or lens implantation. In the case of hospital bills, records that indicate a primary diagnosis of cataract (ICD-9 366, 366.1, 366.9) with a charge of more than $400 also will be accepted as confirmatory evidence.

B. A paid anesthesiologist's bill coded for cataract surgery (Table 1) or using the anesthesia code for lens surgery (CPT 00142)

C. A paid assisting surgeon's bill for cataract surgery (hereafter called "assistant's bill") 2. Primary Evidence: A paid facility bill that indicates cataract surgery (diagnosis alone not

acceptable) plus confirmatory data consisting of either or both of the follOwing: A. A paid assistant's bill for cataract surgery B. A paid anesthesia bill indicating cataract surgery (code for lens procedure is insufficient)

Level II: Must Meet Either Criterion 1, 2, 3, or 4 1. Primary Evidence: A paid surgeon's bill for cataract surgery (without confirmatory evidence) 2. Primary Evidence: A paid assistant's bill for cataract surgery plus confirmatory evidence of a paid

anesthesia bill coded for cataract surgery or that uses the anesthesia codes for lens surgery 3. Primary Evidence: Either:

A. A paid assistant's bill for cataract surgery B. A paid anesthesia bill for cataract surgery (anesthesia code for lens procedure is not

acceptable), with confirmatory evidence of a paid facility bill with a procedure code for lens implantation or lacking a procedure code but with a diagnosis of cataract (ICD-9 366, 366.1, 366.9) and reimbursement of more than $400

4. Primary Evidence: A paid facility bill that indicates cataract surgery with confirmatory evidence of a paid anesthesia bill with a code for lens procedure

Level III: Must Meet Criterion 1, 2, or 3 1. An isolated paid assistant's bill for cataract surgery 2. An isolated paid anesthesia bill for cataract surgery (code for lens procedure is not acceptable) 3. An isolated paid facility bill with procedure code for cataract surgery

coverage during 1986-1988, as well as 1616 persons who were members of Health Maintenance Organizations because complete utilization data were not available on those beneficiaries. Beneficiaries residing outside of the 50 states and the District of Columbia (554 persons) were excluded because they are likely to receive care outside the Medicare program. We also excluded 1511 beneficiaries younger than age 65, because their Medicare entitlement was based on having a chronic disability or having end-stage renal disease, as well as 3061 individuals who were 65 at the time of their cataract surgery because no Medicare data for the year before cataract surgery were available. Because some individuals were excluded for multiple reasons, the total number of exclusions (8300) exceeds the number of persons excluded (8188).

surgery cases included in this analysis. Individuals who underwent intracapsular cataract extraction (n = 8495) or for whom the type of extraction was not specified (n = 2223) were excluded from this analysis. Persons who underwent cataract surgery combined with anterior vitrectomy (n = 554), presumably as a result of intraoperative vitreous loss, were excluded from analysis, since their postoperative experience is likely to differ from that of persons who undergo uncomplicated cataract extraction. Similarly, those whose cataract surgery was combined with corneal, retinal, or glaucoma procedures (n = 1664) were excluded from the study cohort. After these various exclusions, there were a total of 57, 103 level I or level 2 cases of extracapsular cataract extraction left in our sample.

We obtained all Medicare hospital and physician records from 1985-1988 on the remaining 72,111 persons in the cohort, along with demographic data from the Medicare denominator file. We then classified each of the 72, III possible cases of cataract surgery as level I, 2, or 3, based on our level of certainty that cataract surgery occurred (Table I). Classification as a level I case required two pieces of evidence, one of which was a physician's bill.

Table 2 shows the CPT and ICD-9 codes related to extracapsular extraction that were used to identify cataract

We searched the Medicare Part A and B records of these cataract surgery cases through the end of 1988 for evidence of subsequent Nd:YAG capsulotomy, retinal detachment, or retinal break. Most cases of presumed Nd: Y AG capsulotomy were ascertained based on an ophthalmologist's bill (provider type 17, 18, or 49) with CPT code 66821 or a local code for Nd: Y AG capsulotomy. An additional 239 cases of Nd:YAG capsulotomy were ascertained on the basis of either an ASC bill with CPT code 66821 and a provider type of 17 or 18 (ophthalmologist) or on the basis of a hospital outpatient bill with

1489

Ophthalmology Volum-e 99, Number 10, October 1992

Table 2. Codes Used to Ascertain Cases of Extracapsular Cataract Extraction

CPT Codes 66840 Removal of lens material; aspiration technique, one or

more stages 66915 Expression of lens, linear, one or more stages 66940 Extraction of lens with or without iridectomy;

extracapsular 66984 Extracapsular cataract removal with insertion of

intraocular lens prosthesis (one stage procedure), manual or phacoemulsification technique

66850 Removal of lens material; phacofragmentation technique (mechanical or ultrasonic, e.g., phacoemulsification), with aspiration

ICD-9 Procedure Codes 13.2 Extracapsular extraction of lens by linear extraction

technique 13.3 Extracapsular extraction of lens by simple aspiration

(and itrigation) technique; irrigation of traumati:c cataract

13.4 Extracapsular extraction of lens by fragmentation and aspiration technique

13.41 Phacoemulsification and aspiration of cataract 13.43 Mechanical phacoemulsification and other aspiration

of cataract 13.5 Other extracapsular extraction oflens 13.51 Extracapsular extraction oflens by temporal inferior

route 13.59 Other extracapsular extraction of lens

CPT code 66821 (no provider type is specified in this file). Possible episodes of retinal detachment were classified as level I or 2 in a manner similar to that employed for cases of cataract surgery (Table 3).

Statistical Methods

We used standard actuarial lifetable analysis and Cox's proportional hazards modef3 in our analyses. All patients were entered into analysis at the time of cataract surgery. Nd:YAG laser capsulotomy was treated as a time-dependent covariate, so that the precapsulotomy experience would be counted in the noncapsulotomy group and the postcapsulotomy experience would be counted in the capsulotomy group. Cases were withdrawn from observation (censored) either: (1) at the end of 1988 (48,827 cases), (2) on the date of death (5791 cases), (3) on detection of subsequent intraocular surgery other than cataract extraction or an episode of ocular trauma (1674 cases), or (4) when beneficiaries joined an HMO, lost Medicare coverage, or moved out of the United States (817 cases). Data were analyzed using STAT A (Computing Resource Center, Santa Monica, CA) and SAS (SAS Institute, Cary, NC) on a Solbourne model 5E/803 computer (Solbourne Inc, Longmount, CO).

1490

Results

Using the algorithms defined above, we identified 57,103 level 1 or level 2 cases of extracapsular (either nuclear expression or phacoemulsification) cataract surgery from the 1986-1987 Medicare sample (Table 4). This cohort of patients, whose demographic and clinical characteristics are listed in Table 5, constituted our primary analysis group and received an average of 24 months of followup (range, 12 to 36 months) for a total of 1.3 million person-months of observation.

Of all persons who underwent level I or level 2 cataract surgery, 13,709 (24%) satisfied our criteria for having undergone Nd:YAG laser capsulotomy during the followup period. In 44 of these cases, the date listed for capsulotomy was the same as that for cataract surgery. Although these cases may represent anterior capsulotomy, we chose to leave them in the analysis.

The rate of capsulotomy per person-year of presumed pseudophakia varied considerably by state, with a range from 2.5% in Wyoming to 27% in Nevada (coefficient of variation, 33.77) (Fig 2). When states in which fewer than 100 capsulotomies were performed during the period of observation are eliminated from consideration, there is a 2.5-fold variation observed in rate of capsulotomy (range, 10% to 24%). There also is considerable variation in the timing of posterior capsulotomy after cataract extraction among patients who undergo capsulotomy (Fig 3). Over 2 years of observation, the median time to capsulotomy is 363 days (lower quartile, 154 days; upper quartile, 555 days).

A total of 337 persons were identified with possible episodes of retinal detachment and an additional 194 with repair of a retinal break. None of the 44 individuals for whom the listed date of capsulotomy was the same as that listed for surgery (possible cases of anterior capsulotomy) subsequently experienced a retinal break or detachment. Similarly, none of the 239 patients in whom capsulotomy was ascertained solely on the basis of a facility bill subsequently experienced a retinal detachment. In analyzing risk factors for retinal complications, cases of levelland level 2 cataract surgery and level 1 retinal detachment were included in our primary analysis. All other combinations oflevel of cataract surgery, retinal break, and retinal detachment were considered in sensitivity analyses.

By lifetable analysis, the retinal detachment rate at 3 years after Nd:YAG capsulotomy is 1.2% (Fig 4). However, when capsulotomy is chosen as the starting time, there is no appropriate time axis on which to plot a comparison (noncapsulotomy) group. Moreover, patients who undergo posterior capsulotomy have, by necessity, first undergone cataract extraction. Therefore, the probability of retinal detachment is influenced by both the hazard attributable to cataract surgery and that attributable to capsulotomy. To distinguish between these two risks, we treated capsulotomy as a time-dependent covariate in the analyses below.

By lifetable analysis from time of initial cataract surgery, the cumulative probability of retinal detachment

Table 3. Classification of Possible Episodes of Retinal Complications

Retinal Detachment Level I (Most Stringent) Must Meet Either Criterion I, 2, or 3: 1. Primary Evidence: A paid surgeon's bill for repair of retinal detachment plus one or more of

the following pieces of confirmatory data: A. A paid bill or discharge record from an ambulatory surgery center, hospital outpatient

department or short stay hospital indicating either repair of retinal detachment or principal diagnosis of retinal detachment.

B. A paid anesthesiologist's bill coded for repair of retinal detachment (hereafter called "anesthesia bill")

C. A paid assisting surgeon's bill for repair of retinal detachment (hereafter called "assistant's bUI")

2. Primary Evidence: A paid facility bill that indicates both repair of retinal detachment and principal or secondary diagnosis of retinal detachment (no confirmatory data needed)

3. Primary Evidence: A paid facUity bill that indicates repair of retinal detachment or coded with principal diagnosis of retinal detachment plus confirmatory data consisting of either or both of the following: A. A paid assistant's bill for repair of retinal detachment B. A paid anesthesia bill indicating repair of retinal detachment

Level II: Must Meet Either Criterion I, 2,3,4, or 5: 1.. Primary Evidence: A paid surgeon's bill for repair of retinal detachment (without

confirmatory evidence) 2. Primary Evidence: A paid hospital inpatient record that indicates repair of retinal detachment. 3. Primary Evidence: A paid hospital inpatient record that indicates principal diagnosis of retinal

detachment. 4. Primary Evidence: A paid assistant's bill for repair of retinal detachment plus confirmatory

evidence of a paid anesthesia bill coded for repair of retinal detachment. 5. Primary Evidence: Either (1) a paid assistant's bill for repair of retinal detachment or (2) a paid

anesthesia bill for repair of retinal detachment, with confirmatory evidence of a paid facility bill with a secondary diagnosis of retinal detachment.

Level III: Must Meet Criterion I, 2, 3, or 4: 1. An isolated paid assistant's bill for repair of retinal detachment 2. An isolated paid ASC bill for repair of retinal detachment 3. An isolated paid anesthesia bill for repair of retinal detachment 4. An isolated paid facility bill with secondary diagnosis of repair of retinal detachment

Retinal Break Level I (Most Stringent) Must Meet Either Criterion I, 2, or 3: 1. Primary Evidence: A paid surgeon's bill for prophylaxis of retinal detachment (no

confirmatory data needed) 2. Primary Evidence: A paid facility bill that indicates both prophylaxis of retinal detachment

and principal or secondary diagnosis of retinal break (no confirmatory data needed) 3. Primary Evidence: A paid facility bill that indicates prophylaxis of retinal detachment or

coded with principal diagnosis of retinal break plus confirmatory data consisting of either or both of the follOwing: A. A paid assistant's bill for prophylaxis of retinal detachment B. A paid anesthesia bill indicating prophylaxis of retinal detachment

Level II: Must Meet Either Criterion I, 2, or 3: 1. Primary Evidence: A paid inpatient hospital bill that indicates prophylaxis of retinal

detachment or primary diagnosis of retinal break 2. Primary Evidence: A paid assistant's bill for prophylaxis of retinal detachment plus

confirmatory evidence of a paid anesthesia bill coded for prophylaxis of retinal detachment 3. Primary Evidence: Either 1) a paid assistant's bill for prophylaxis of retinal detachment or 2) a

paid anesthesia bill for prophylaxis of retinal detachment, with confirmatory evidence of a paid facility bill with a secondary diagnosis of retinal break

Level III: Must Meet Criterion I, 2, 3, 4, or 5: 1. An isolated paid assistant's bill for prophylaxis of retinal detachment 2. An isolated paid ASC bill for prophylaxis of retinal detachment 3. An isolated paid anesthesia bill for prophylaxis of retinal detachment 4. An isolated paid inpatient hospital bill with secondary diagnosis of retinal break 5. An isolated paid outpatient hospital bill with principal diagnosis of retinal break

1491


Table 4. Type of Record Used to Ascertain First Cataract Surgery in 1986-87

Levell: Basis for inclusion Surgeon + facility:

plus Anesthesia: Assistant: Anesthesia assistant: Surgeon + anesthesia: Surgeon + assistant: Surgeon + anesthesia and assistant: Facility + anesthesia: Facility + assistant: Facility + anesthesia and assistant: Total, Levell:

Level 2: Basis for inclusion Surgeon only: Anesthesia + assistant: Facility + anesthesia: Facility + assistant: Facility + anesthesia and assistant: Misc. surgeon + (facility, anesthesia, or

assistant): Total, Level 2:

9121

28,520 445

2685 8801

119 649

2635 17

194 53,186

1754 26

427 3

57

1650 3917

(level 1) over the first 36 months following cataract extraction (level 1 or 2) was greater in patients who underwent capsulotomy than in those who underwent cataract extraction alone (1.6% versus 0.8%; P < 0.02) (Fig 5A). A similar relationship (Fig 5B) is observed in the cumulative probability oflevell or 2 retinal detachment (1.9% versus 1.0%; P < 0.01). In these analyses, patients may cross from the noncapsulotomy to the capsulotomy group at time of capsulotomy, so that person-days of observation before capsulotomy are counted in the noncapsulotomy group and person-days of observation after capsulotomy are counted in the capsulotomy group. Examination of the underlying hazard function (i.e., incidence of retinal detachment per month after cataract surgery) shows that, in both cases, the curves are roughly proportional and do not cross, satisfying the major preconditions of the Cox model (Figs 6A and 6B).

Cox's proportional hazard model, based on levelland 2 cataract surgery and level 1 retinal detachment suggests that (Table 6) performance of capsulotomy, younger age, white race, and male sex, in order of decreasing importance, are associated with a significant increase in the risk of retinal detachment, whereas a diagnosis of diabetes mellitus, performance of bilateral cataract surgery, and implantation of an intraocular lens are not associated with an increase in the risk of retinal detachment. As shown in Table 6, the relative risk of retinal detachment in the capsulotomy group compared with the noncapsulotomy group is 3.9 (95% confidence interval, 2.89 to 5.25). Compared with persons older than 85 years, those 75 to 84 years of age and 65 to 74 years of age have a 3-fold and 7-fold increased risk of retinal detachment, respectively. When age ranging from 65 to 85 years is considered as a

1492

continuous, rather than categorical variable, each younger year of age below 85 years is associated with a 1.09-fold increase in the risk of retinal detachment. When the outcome of interest is broadened to include retinal break or detachment, a diagnosis of diabetes mellitus is associated with a 2-fold increase in risk, based on the Cox model (data not shown; P < 0.05).

Cox's proportional hazards model also was used to perform sensitivity analyses to assess the effect of definition of cataract surgery and retinal complications on the relative hazard associated with capsulotomy (Table 7). In all analyses, patient age, sex, race, undergoing bilateral cataract surgery between 1985 and 1988, and diagnosis of diabetes were included in the model.

As shown in Table 7, the relative hazard associated with capsulotomy in the primary analysis (levelland 2 cataract surgery and level 1 retinal complication; relative risk, 3.90) is not significantly different from that associated with the most stringent definition of cataract surgery and complications (level 1 cataract surgery versus level 1 retinal detachment; relative risk, 4.10). Nor does considering

Table 5. Demographic and Clinical Characteristics of the Cohort

Age (yrs) No. of Persons Percent of Cohort

66-69 6833 12.0

70-74 13,303 23.3

75-79 15,662 27.4 80-84 12,459 21.8 85+ 8846 15.5

Race No. of Persons Percent of Cohort

White 52,366 91.7 Black 3227 5.7 Other 564 1.0 Unknown 946 1.7

Sex No. of Persons Percent of Cohort

Male 19,524 34.2 Female 37,579 65.8

Intraocular Lens Implanted No. of Persons Percent of Cohort

Yes 55,961 98 No 1142 2

Bilateral Cataract Surgery No. of Persons Percent of Cohort

Yes 17,028 29.8 No 40,075 70.2

Diabetes Mellitus No. of Persons Percent of Cohort

Yes 8276 14.5 No 48,827 85.5

Javitt et al . National Outcomes of Cataract Extraction

Figure 2. Variation in likelihood. of posterior capsulotomy after extracapsular cataract extraction by state per person-year of observation.

the least stringent definition of retinal detachment (level 1 or 2 cataract surgery and either level I or level 2 retinal detachment; relative risk, 3.10) significantly alter the results.

The relative risk of retinal break also is higher in the capsulotomy group than in the noncapsulotomy group (Table 7)_ When level 1 or 2 cataract surgery cases and level 1 retinal break are considered, the relative risk in the capsulotomy versus the noncapsulotomy group is 2.24 95% confidence interval, 1.39 to 3.59). Sensitivity analyses

o 3 6 9 12 15 18 21 24 Months post Surgery

Figure 3. Timing of capsulotomy after cataract extraction over the first 2 years of observation.

Rote o f N d:YAG Ca p sulo tomy

D • • • 0.0 4 t o 0. 1 2

0. 12 t o 0.14

0 .14 to 0.18

0.18 to 0.28

across all definitions of cataract surgery and retinal break give consistent results.

To exclude the possibility that capsulotomy was performed to provide better visibility in repairing a pre-existing retinal detachment, we excluded all cases in which retinal detachment occurred within 0 to 7 days of capsulotomy and found no change in the relative risk of retinal detachment associated with capsulotomy.

Q 0::

1.2')(".---------------;::=====1

1.0% ••• ••••• ••• ••• ••• ••• ••• ••• •• • •• • ••• ••• •• •• •• ••• ••• • ••••••• • •••••• • •• •

'0 0.8% •••••••••••••••••••••••••••• • •••••••••••••••••••••••••••••••••••••••••

. ~ ~ e 0.6% •••••••••••••••••• • •••••••••••••••••••••••••••••••••••••••••••••••••

'" .~ .. ~ 0.4% ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••

a o. ••• • ••••••••••• ••••••••••••••••••••••••••••••••••••••••••••••••••••••

90 160 270 360 450 540 630 720 810 900 990 Days Post Capsu)otomy

Figure 4. Lifetable analysis (actuarial) o f cumulative probability of r etinal detachment (RD) from time of c apsulotomy in those who underwent level 1 or level 2 cataract surgery.

1493


2.0% 2.0%.----------------------,

,," / -"

" "

,--,

,.." .... ___________________ .. ________ .... _ .. _____ -I ___ ........ _c;~~~'!.1?!~~!_ ... ____ _

/ ,,, ~'

" ... "

,,' ,." ,,'

,..'

Logrank test: p < .02

0.0%.lf-I-r-T"T"T""T"""T"'r-r-T1"T"2"T""T"T"""'1r-

6 T""T"""T"20..--r"""""""2r-

4 T""T""T:T::-r-r....."r:-T"T""T"~

A Months post Cataract Surgery

,,/

,," "" ...

,," ,'" Capsulotomy ,

,..'

,-,

.................................. , .............................. . ,.. .....

~ /

'-

,,-" , ,,/ " .....

,,/

O.~k~~~~~~~~~~~~~~Lo~~~ank~Tw~:p~<~.02~~ 4 8 12 16 20 24 28 32 36

B Months post Cataract Surgery

Figure 5. Lifetable analysis (actuarial) of cumulative probability of retinal detachment (RD) after level 1 or level 2 cataract surgery for those who did and did not subsequently undergo capsulotomy. Probability of level 1 detachment is shown in A and probability of level 1 or 2 detachment is shown in B.

Discussion

Our findings suggest that Nd: Y AG laser capsulotomy is associated with approximately a fourfold increase in the risk of aphakic or pseudophakic retinal detachment, and a twofold increase in the risk of retinal break (without detachment), when we compare patients who undergo extracapsular cataract extraction with and without subsequent capsulotomy (Tables 6 and 7). Sensitivity analyses, using alternative ascertainment criteria for cataract surgery and retinal complications, yield consistent estimates of this increased risk.

Multivariate analysis (Table 6) demonstrates that, in addition to capsulotomy, younger patient age, white race, and male sex (in order of decreasing importance) are sig-

0.02 -,-----------------------,

"E a 0.01 l:

0.00 +---,--,-..,.---,--,--,----;-...,--,--;--...,..---1 o 12 24 36

A Months Post Cataract Surgery

nificantly associated with increased risk of retinal complications. These variables also were significantly associated with retinal detachment in our previous analysis of 330,000 inpatient cataract extractions performed during 1984 and are discussed in detail in that report.24 In the context of posterior capsulotomy, the data emphasize that ophthalmologists should consider the increased risk of retinal complications associated with these risk factors in advising patients whether to undergo capsulotomy and in informing them of possible short-term and long-term risks.

We have previously reported a higher risk of retinal detachment in persons who undergo intracapsular cataract extraction, compared with those who undergo extracapsular surgery.24 In that report, we hypothesized that removal of the posterior capsule might predispose to vitreous

0.02 -r------------------,

"E a 0.Q1 l:

Capsulotomy

No Capsulotomy

0.00 -I----,-,...-,---;--,----;-...,----,--T'=-,---r---1 o 12 24 36

B Months Post Cataract Surgery

Figure 6. Hazard (point incidence) of retinal detachment after level 1 or level 2 cataract surgery for those who did and did not subsequently undergo capsulotomy. Hazard of level 1 detachment is shown in A and hazard of level 1 or 2 detachment is shown in B (error bars represent standard error of the mean).

1494


Table 6. Cox Proportional Hazards Model (Based on Levelland 2 Cataract Surgery Cases and Level

1 Retinal Detachment)

(95% Confidence P

Relative Hazard Variable Limits) Value

Posterior Capsulotomy Not performed Reference Performed 3.90 (2.89, 5.29) <0.0001

Age 85 years and over Reference 75-84 years 3.19 (1.55, 6.57) <0.0001 65-74 years 7.20 (3.53, 14.68) <0.0001

Race Black Reference White 3.52 (1.31, 9.48) 0.02

Sex Female Reference Male 1.72 (1.34, 2.21) <0.0001

Diabetes Mellitus Diagnosis absent Reference Diagnosis present 1.18 (0.85, 1.67) NS

Bilateral Cataract Surgery· Absent Reference Present 1.17 (0.91, 1.51) NS

Intraocular Lens Implanted Reference Not implanted 0.92 (0.38, 2.23) NS

• During observation period.

collapse and increased vitreous traction. The analyses presented in this report are consistent with that hypothesis.

Substantial variation in timing and utilization of surgical procedures may be a function of provider uncertainty as to the most appropriate indications for those procedures. The variation in utilization of capsulotomy across large regions suggests a lack of consensus on the appropriate timing and indications for capsulotomy. Wennberg25 and others26 have shown that variation in

utilization of surgical procedures is more appropriately studied in small geographic regions. Analysis of variation in capsulotomy by patient and regional characteristics will be presented in future reports.

Interpretation of our findings requires understanding of the biases we have previously discussed that are inherent in all Medicare data base analyses.24,27-29 Of principal concern in any analysis based on claims data are possible errors in identifying cases (in this study, cases of cataract surgery) and complications. Fisher and co-workers3o have shown that Medicare hospital records are 95% accurate in the coding of ophthalmologic procedures. No validation study has ever been performed on the coding of ophthalmologic procedures in Medicare physicians' bill (part B) data. However, our sensitivity analyses demonstrate consistent results despite varying the criteria for case and complication ascertainment (Table 7), thereby making our conclusions more robust.

Because laterality is only coded in approximately 10% of Medicare records, we cannot distinguish in which eye the cataract surgery, the capsulotomy, or a retinal detachment or break occurred and, therefore, cannot be certain that any retinal complication that we ascertained from Medicare claims occurred in the same eye that underwent cataract surgery or capsulotomy. Unless there is some reason why capsulotomy is more likely to be performed in patients who are at increased risk of having a future retinal complication in the other eye, however, there is no reason to believe that this ambiguity introduces bias into our analysis. Moreover, the underlying rate of retinal detachment repair in the entire Medicare population is approximately 0.1 %, an order of magnitude lower than the rate even in our noncapsulotomy group.

Nonetheless, because of the lack of information in Medicare data regarding which eye underwent cataract surgery, capsulotomy, and repair of a retinal detachment or break, the association between capsulotomy and risk of retinal detachment that we have identified should be considered probable rather than definite. The only way to validate our findings would be through review of the clinical records of individuals included in our analysis of Medicare claims data, a study that we believe should be undertaken.

Table 7. Relative Hazard of Retinal Break and Retinal Detachment Associated with Capsulotomy

Level I Level I or II Complication Cataract Surgery Cataract Surgery

Retinal Detachment ReI Hazard (95%CL) Rel Hazard (95%CL) Level I 4.10 (3.01, 5.58) 3.90 (2.89, 5,25) Level I or II 3,22 (2.43, 4.27) 3.10 (2.36, 4.07)

Retinal Break Level I 1.94 (1.21,3.12) 2.24 (1.39, 3.59) Level I or II 2.26 (1.49, 3.43) 2.49 (1.68, 3.69)

Retinal Break or Detachment Level I 3.03 (2.33, 3.95) 3.06 (2.38, 5.95) Level I or II 2.53 (1.97, 3.26) 2.48 (1.95,3.16)

1495


Our analysis may have underestimated the increased risk of retinal complications associated with capsulotomy for two reasons. First, we had no way of identifying patients who underwent a primary incisional capsulotomy (planned or not) at the time of cataract surgery. Therefore, these patients and their retinal complications are included in the noncapsulotomy group. If anything, this caused a reduction in the relative risk of retinal detachment associated with capsulotomy.

Second, vitreous loss may have occurred at the time of cataract surgery but may not have been reported. Previously, we have reported that vitreous loss with concomitant anterior vitrectomy is associated with a 4.5-fold increased incidence of retinal detachment after cataract surgery?4 Although we have eliminated from consideration those persons for whom vitrectomy was billed at the time of cataract surgery, those patients for whom vitrectomy was not billed are included among the noncapsulotomy group. Again, the effect is to underestimate the risk of retinal detachment associated with capsulotomy.

We considered the possibility that, in some cases, capsulotomy may be performed immediately before repair of an existing retinal detachment to allow for better visualization of the posterior segment. The exclusion of all cases of retinal detachment occurring within 1 to 7 days of capsulotomy had no effect on our findings.

We followed patients in the Medicare data base for an average of 24 months (range, 12 to 36 months) after cataract surgery. It is possible that a longer period of observation would yield different findings. We will be able to examine this possibility in the future.

Although diabetes was not a significant risk factor in most of our multivariate models, we are limited in our ability to study diabetes as a risk factor. Diabetes was only identified when it was listed as a primary or secondary diagnosis on a hospitalization during the 1985 to 1988 period of observation. The National Diabetes Data Group has reported that many hospital discharge records for diabetic persons do not list diabetes as a diagnosis (personal communication; Maureen Harris, National Diabetes Data Group).

Retrospective studies have suggested that lattice degeneration with holes, high myopia, larger capsulotomy size, and increased laser energy may predispose patients to retinal complications after Nd:YAG capsulotomy.31 None of these risk factors can be ascertained from the Medicare claims data base.

The demonstration of a statistically significant association does not definitively prove that Nd:YAG capsulotomy causes retinal detachment. Possibly, there are risk factors that cause posterior capsular opacification and also cause retinal detachment. Because patients with significant posterior capsular opacity are most likely to undergo capsulotomy, an exogenous factor that increased the incidence of both posterior capsular opacification and retinal detachment would lead to a statistically significant, although nonetiologic, association. However, no such factor sufficiently prevalent to explain this association has been described.

1496

A prospective cohort study or case-control study in a large and representative group of patients is needed to better understand those factors that may predispose one to retinal complications after cataract surgery in general and after capsulotomy in particular. If there are clinical factors predisposing to retinal complications that we have not measured, a prospective or case-control study would be necessary to estimate the risk associated with those factors.

Conclusions

Although Nd:Y AG capsulotomy remains the treatment of choice for patients with clinically significant posterior capsular opacity (i.e., patients with a decrease in visual function, functional status, or quality of life), the increased risk of retinal complications that is associated with capsulotomy should be considered by physicians and their patients in choosing whether to use the procedure. Younger patients and those with clinical risk factors that predispose to retinal detachment are at particular risk.

Given the now documented and previously suspected risk of retinal complications, it would be prudent to defer capsulotomy until the patient's impairment due to capsular opacification warrants the increased risk of retinal complications associated with capsulotomy.

Appendix

Cataract Patient Outcomes Research Team: Earl P. Steinberg, MD, MPP (principal investigator), Marilyn Bergner, PhD, (co-principal investigator), Alfred Sommer, MD, MHS (co-principal investigator), Gerard F. Anderson, PhD, Eric B. Bass, MD, MPH, Joseph Canner, MHS, Alan M. Gittelsohn, PhD, Jonathan C. Javitt, MD, MPH, Margaret M. Kolb, DrPH, Marcia Legro, ScD, Neil R. Powe, MD, Oliver P. Schein, MD, MPH, Phoebe Sharkey, ScD, Donald M. Steinwachs, SeD, Debra Street, MPH, and James M. Tielsch, PhD. The authors wish to acknowledge the members of the Advisory Board of the Cataract Patient Outcomes Research Team: Donald Doughman, MD, Merton Flom, OD, Thomas Harbin, MD, Harry Knopf, MD, Thomas Lewis, OD, Stephen Obstbaum, MD, Denis O'Day, MD, Walter Stark, MD, Arlo Terry, MD, and C. Pat Wilkinson, MD.

References

1. Wilhelmus KR, Emery JM. Posterior capsule opacification following phacoemulsification. Ophthalmic Surg 1980;11: 264-7.

2. Sterling S, Wood TO. Effect of intraocular lens convexity on posterior capsule opacification. J Cataract Refract Surg 1986;12:655-7.

3. Aron-Rosa D, Grieseman J-C, Aron J-J. Use of a pulsed neodymium Y AG laser (picosecond) to open the posterior lens capsule in traumatic cataract: a preliminary report. Ophthalmic Surg 1981; 12:496-9.


4. Steinert RF, Puliafito CA, Kumar SR, et al. Cystoid macular edema, retinal detachment, and glaucoma after Nd:YAG laser posterior capsulotomy. Am J Ophthalmol 1991;112: 373-80.

5. Aron-Rosa DS, Aron J-J, Cohn HC. Use of a pulsed picosecond ND:YAG laser in 6,664 cases. JAm Intraocul Implant Soc 1984;10:35-9.

6. Keates RH, Steinert RF, Puliafito CA, Maxwell SK. Longterm follow-up ofND:YAG laser posterior capsulotomy. J Am Intraocul Implant Soc 1984;10:164-8.

7. Johnson SH, Kratz RP, Olson PF. Clinical experience with the Nd:Y AG laser. J Am Intraocul Implant Soc 1984; 10: 452-60.

8. Chambless WS. Neodymium:YAG laser posterior capsulotomy results and complications. J Am Intraocul Implant Soc 1985;11:31-2.

9. Stark WJ, Worthen D, Holladay JT, Murray G. Neodymium:YAG lasers. An FDA report. Ophthalmology 1985;92: 209-12.

10. Winslow RL, Taylor Be. Retinal complications following YAG capsulotomy. Ophthalmology 1985;92:785-9.

11. Liesegang TJ, Bourne WM, Ilstrup DM. Secondary surgical and neodymium-Y AG discissions. Am J Ophthalmol 1985; 100:510-9.

12. Vester CAGM, Bienfait MF, de Jong PTVM, Pameijer JH. Retinal detachment following neodymium-Y AG laser capsulotomy. Fortschr Ophthalmol 1986;83:441-3.

13. Shah GR, Gills JP, Durham DG, Ausmus WHo Three thousand Y AG lasers in posterior capsulotomies: an analysis of complications and comparison to polishing and surgical discission. Ophthalmic Surg 1986;17:473-7.

14. Bath PE, Fankhauser F. Long-term results ofNd:YAG laser posterior capsulotomy with the Swiss laser. J Cataract Refract Surg 1986;12:150-3.

15. Ficker LA, Vickers S, Capon MRC, et al. Retinal detachment following Nd:YAG posterior capsulotomy. Eye 1987;1: 86-9.

16. Ambler JS, Constable 11. Retinal detachment following Nd: YAG capsulotomy. Aust N Z J OphthalmoI1988;16:337-41.

17. Rickman-Barger L, Horine CW, Larson RS, Lindstrom RL. Retinal detachment after neodymium:YAG laser posterior capsulotomy. Am J Ophthalmol 1989;107:531-6.

18. Dardenne MU, Gerten GJ, Kokkas K, Kermani O. Retrospective study of retinal detachment following neodymium:YAG laser posterior capsulotomy. J Cataract Refract Surg 1989; 15:676-80.

19. Lave J, Dobson A, Walton e. The potential use of Health Care Financing Administration data sets for health care services research. Health Care Financ Rev 1983;5:93-8.

20. The International Classification of Diseases: 9th revision, Clinical Modification. ICD-9-CM, 3rd ed. Vol. 1: Diseases, Tabular List. [Washington DC]: US Dept. Health Human Services, 1989.

21. HCFA Common Procedure Coding System (HCPCS), 1989 (non-CPT-4). [Baltimore]: Health Care Financing Administration, 1989.

22. Physicians' Current Procedural Terminology: CPT, 4th ed. Chicago: American Medical Association, 1977.

23. Cox DR. Regression models and life-tables. J R Stat Soc [B] 1972;34: 187-220.

24. Javitt JC, Vitale S, Canner JK, et al. National outcomes of cataract extraction. I. Retinal detachment after inpatient surgery. Ophthalmology 1991 ;98:895-902.

25. Wennberg IE. Population illness rates do not explain population hospitalization rates. A comment on Mark Blumberg's thesis that morbidity adjusters are needed to interpret small area variations. Med Care 1987;25:354-9.

26. Chassin MR, Kosecoff J, Park RE, et al. Does inappropriate use explain geographic variations in the use of health care services? A study of three procedures. JAMA 1987;258: 2533-7.

27. Steinberg EP, Whittle J, Anderson GF. Impact of claims data research on clinical practice. Int J Technol Assess Health Care 1990;6:282-7.

28. Javitt Je. Outcomes of eye care from Medicare data [editorial]. Arch Ophthalmol 1991; 109: 1079-80.

29. Javitt JC, Vitale S, Canner JK, et al. National outcomes of cataract extraction. Endophthalmitis following inpatient surgery. Arch Ophthalmol 1991; 1 09: 1085-9.

30. Fisher ES, Whaley FS, Krushat WM, et al. The accuracy of Medicare's hospital claims data: progress has been made, but problems remain. Am J Public Health 1992;82:243-8.

31. Koch DD, Liu JF, Gill EP, Parke DW II. Axial myopia increases the risk of retinal complications after neodymiumY AG laser posterior capsulotomy. Arch Ophthalmol 1989; 1 07:986-90.

Discussion by

C. P. Wilkinson, MD In the United States, more than one million posterior chamber lenses have been implanted during each of the last 4 years (Stark and Wilkinson; unpublished FDA data obtained from IOL manufacturers, 1991). Therefore, more than 10,000 patients will be annually affected by any complication of cataract surgery that occurs in only 1 % of cases. Retinal detachment is probably the most common potentially blinding complication that occurs after cataract surgery, and this report clearly indicates that YAG capsulotomies significantly increase the risk of subsequent retinal tears and detachments.

From the Dean A. McGee Eye Institute and the Department of Ophthalmology, University of Oklahoma Health Sciences Center, Oklahoma City.

Javitt and co-workers have used an immense unselected data base in their evaluation of retinal complications of Y AG capsulotomy. A randomly selected cohort of 46,000 individuals undergoing cataract surgery in 1986 and 1987 were followed for a minimum of 1 year, and between a 3-fold and 4-fold increase in risk of retinal detachment was demonstrated in eyes in which a Y AG capsulotomy had been performed. Retinal detachments occur more commonly in eyes with so-called "high-risk features," such as high myopia, lattice degeneration, and a history of detachment in the fellow eye, and this is true regardless of whether the lens capsule is intact. Since the incidence of capsular opacification does not appear to be associated with these high-risk features, they were presumably present in equal percentages of eyes undergoing capsulotomy and in those in which the capsules remained intact. It is interesting that these data are remarkably

1497


consistent with a recent retrospective study of Kraff and Sanders,2

who showed that eyes undergoing primary capsulotomy had approximately a threefold greater incidence of subsequent detachment, whether or not the additional high-risk feature of high myopia was present. In an article published last year, Javitt et al3 demonstrated that the risk of rehospitalization for retinal detachment was more than 50% greater after intracapsular than after extracapsular surgery. This current report emphasizes that the protective effect of extracapsular procedures is significantly negated by opening the posterior capsule. Although at this time we do not know if the rate of detachment in these eyes is equal to that after intracapsular surgery, it does appear to be approaching it.

The precise reasons that primary and secondary capsulotomies are associated with an increased rate of retinal tears and detachments remains unknown, but alterations in the vitreous gel are typically responsible for these problems, and it is quite likely that the opening of the posterior lens capsule causes structural changes in this important substance. Deliberate defocusing of the Y AG laser beam probably causes small ruptures in the anterior hyaloid face in most cases, and more significant vitreous damage may occur, even if we can not yet recognize it. The most important structural change that has been demonstrated to date in eyes at autopsy is the loss of hyaluronic acid from the vitreous gel after intracapsular removal of the crystalline lens or after the opening of the posterior lens capsule. Loss of this component of the gel presumably leads to increased vitreous instability and a greater likelihood of posterior vitreous detachment (PVD). McDonnell et al4 observed that the incidence ofPVD in aphakic and pseudophakic eyes at autopsy was comparable in intracapsular and open-capsule extracapsular cases. However, PVD was approximately 50% less common in capsule-intact cases. Posterior vitreous detachment is the pivotal event usually responsible for the creation of retinal tears and subsequent detachments. When PVD occurs, traction-induced retinal tears develop, and these are more likely in eyes with additional high-risk features, such as high myopia and lattice degeneration. Alternatively, if retinal tears do not occur at the time ofPVD, eyes appear to be at least relatively "protected" from the development of future retinal breaks. Hovland5 studied aphakic fellow eyes of patients with retinal detachment in one eye, a genuine high-risk group

1498

of cases, and demonstrated a lO-fold increase in the incidence of retinal detachment in eyes in which a PVD had not occurred when these were compared with cases in which this event had already taken place. Because PVDs are relatively uncommon in phakic eyes before the eighth decade oflife,6 it is not surprising that relatively youthful eyes undergoing cataract surgery and Y AG capsulotomy appear to be at relatively high risk for the development of retinal detachment.

Clearly, YAG capsulotomies are not innocuous or inexpensive procedures, particularly in cases that exhibit high-risk features. The rather surprising geographical differences in the frequencies in which the procedure is used surely will attract the attention of those currently assessing outcomes after cataract surgery. At the present time, it is appropriate that patients undergo Y AG capsulotomy only when significant symptomatology is present, and a complete informed consent should be obtained. In addition, all such patients should be advised of the classic symptoms of posterior vitreous separation, for these usually herald the onset of a retinal tear, and reattachment success rates are higher when cases are detected relatively early.

References

I. Michels RG, Wilkinson CP, Rice T A. Retinal Detachment. St. Louis: CV Mosby, 1990; 172-91.

2. Kraff MC, Sanders DR. Incidence of retinal detachment following posterior chamber intraocular lens surgery. J Cat Refract Surg 1990;16:477-80.

3. Javitt JC, Vitale S, Canner JK, et al. National outcomes of cataract extraction. I. Retinal detachment after inpatient surgery. Ophthalmology 1991;98:895-902.

4. McDonnell PJ, Patel A, Green WR. Comparison of intracapsular and extracapsular surgery. Histopathologic study of eyes obtained postmortem. Ophthalmology 1985;92: 1208-23.

5. Hovland KR. Vitreous findings in fellow eyes of aphakic retinal detachment. Am J Ophthalmol 1978;86:350-3.

6. Foos RY, Wheeler NC. Vitreoretinal juncture. Synchesis senilis and posterior vitreous detachment. Ophthalmology 1982;89: 1502-12.

Documents

National Outcomes of Cataract Extraction