This study was conducted at University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, USA, and was approved by the Institutional review board. The participants and their parents gave consent. We reviewed the clinical records over 25 years of all children in our academic pediatric ophthalmology practice that developed glaucoma following cataract surgery between 1982 and 2000. All cataracts were identified before one year of age and removed before 5 years of age. All operations were performed by ophthalmic surgeon A.W.B. Patients were excluded if they had: (1) a history of increased intraocular pressure (IOP) before cataract surgery, (2) signs of congenital glaucoma, (3) conditions other than cataract that might predispose the development of glaucoma such as rubella or persistent hyperplasic primary vitreous (PHPV), (4) syndrome-associated cataracts or (5) cataracts related to ocular trauma.

A diagnosis of glaucoma was established by having at least two of the following criteria: (1) repeated IOP measurements over 21 mmHg, (2) cupping of the optic nerve consistent with glaucoma and (3) visual field loss. The following data were recorded: cataract morphology, age at cataract surgery, corneal diameter, central corneal thickness, laterality of glaucoma, surgical techniques used to remove cataract, number and type of secondary procedures, primary or secondary insertion of an intraocular lens (IOL), visual fields, age at glaucoma diagnosis, and histopathologic findings of one eye that was enucleated. The patient with the enucleated eye was a 15-year-old female patient with uncontrollable elevation of intraocular pressure resulting in a painful, blind eye that severely interfered with her life. Corneal diameter was measured with calipers in the operating room during cataract surgery or in the outpatient clinic upon admission for surgery. Central corneal thickness was measured at the same time as corneal diameter. Types and effectiveness of glaucoma therapy were evaluated. Chi square and Mann-Whitney U tests were used for statistical analysis.

Twenty-four patients (37 eyes) were identified with glaucoma after cataract surgery. Patient characteristics are summarized in Tables 1 and 2. All 24 patients were Caucasian (non-Hispanic). Eleven patients had unilateral cataract surgery and unilateral glaucoma and 13 had bilateral cataract surgery and bilateral glaucoma. Follow-up after glaucoma onset ranged from 1 to 23.7 years.

The onset of glaucoma after cataract surgery followed a bimodal pattern (Fig. 1). Fifteen eyes had an early onset of glaucoma (1 week to 13 months after cataract surgery), while 22 eyes had delayed-onset glaucoma (70 to 177 months after cataract surgery). Two patients with bilateral aphakia developed early-onset glaucoma due to pupillary block in the first eye and delayed-onset open angle glaucoma in the fellow eye. Out of 11 patients in the early-onset group, 5 patients had unilateral cataract surgery, 4 patients had bilateral cataract surgery, and 2 patients with bilateral aphakia developed early-onset glaucoma in one eye and delayed-onset glaucoma in the other eye. Out of 15 patients in the delayed-onset group, 6 patients had unilateral cataract surgery, 7 patients had bilateral cataract surgery, and 2 patients with bilateral aphakia developed delayed-onset glaucoma following early-onset glaucoma in the other eye. Early-onset glaucoma typically had an abrupt onset. Presenting symptoms and signs included pain, photophobia, lacrimation, pupillary block, angle closure, rapid myopic shift in the refractive error and corneal edema. Delayed-onset glaucoma (in 18 of 22 eyes) had an insidious, painless onset with an increased IOP that was discovered on routine examination. Age at cataract surgery and type of glaucoma onset were not significantly related (p=0.08).

Corneal diameters were classified as either microcornea or normal-size for the child's age by comparing them to Wallace and Plager's16 table of age-adjusted normal corneal diameters in children. Twelve of 37 (32%) eyes were found to have microcornea, 1 in the early-onset group and 11 in the delayed-onset group. Microcornea was highly associated with delayed-onset open-angle glaucoma (p=0.006, Chi-square).

Lens morphology data were available for 30 eyes. Nuclear cataracts were the most common, occurring in 12 eyes that were equally distributed in the early-onset (33%) and delayed-onset (32%) groups. The remaining cataracts were cortical, lamellar, anterior polar or posterior polar. There was no significant difference in cataract type between eyes that developed early-onset and those that developed delayed-onset glaucoma (p=0.509, Chi-square).

Cataracts were surgically removed by aspiration of the lens material in 20 eyes (9 in the early-onset group and 11 in the delayed-onset group) and aspiration combined with posterior capsulectomy and anterior vitrectomy in 17 eyes (6 in the early-onset group and 11 in the delayed-onset group). Patients who had simple aspiration of the lens material more often required secondary procedures (19 of the 20 eyes) to treat posterior capsule opacification, retained lens material, or pupillary block than patients who also had primary posterior capsulectomy and anterior vitrectomy (3 of the 17 eyes) (p<0.001, Chi-square). A peripheral iridectomy or iridotomy was performed in only 5 eyes during cataract surgery (2 in eyes in early-onset group and 3 in the delayed-onset group). Pupillary block did not occur in any of them. However, in eyes that did not have a peripheral iridectomy, pupillary block was a frequent reason for early-onset glaucoma (8/13) compared to those with delayed-onset of glaucoma (3/19) (p=0.009, Chi-square).

Thirty-three of 37 eyes underwent at least one gonioscopic examination and the filtration angle was closed in at least three quadrants at the time of glaucoma diagnosis in 10 eyes (7 in the early-onset group and 3 in the delayed-onset group, p=0.018, Chi-square). The most consistent gonioscopic findings in the remaining 23 eyes were localized peripheral anterior synechiae (PAS) over a few clock hours, flat iris insertion, poorly developed angle recess, a sheet of translucent tissue overlying the meshwork and increased pigmentation over the trabecular meshwork, narrow or invisible ciliary band, an indistinct scleral spur, and increased vasculature, sometimes bridging the angle (Fig. 2).

Satisfactory reduction in IOP was achieved with one or two topical medications in 2 eyes in the early-onset group and 16 eyes in the delayed-onset group (p<0.001, Chi-square). Surgery was required to manage glaucoma in only 27% (6/22) of eyes in the delayed-onset group but in 87% (13/15) in the early-onset group (p=0.001, Chi-square). Surgical procedures used to control glaucoma in the two groups are summarized in Table 3. Peripheral iridectomy was performed in 5 eyes in the early-onset group, but only 1 eye had a favorable response to this procedure; the others required two or more glaucoma procedures to control the IOP.

Figure 3 shows an angle in the eye of a patient with delayed-onset glaucoma; the meshwork is compact and hypercellular, the iris stroma and neovascular tissue encroaches on the partially developed angle recess and the lumen of Schlemm's canal cannot be clearly identified.

To learn more about the factors specific to the development of glaucoma after cataract surgery, we studied a select group of children who had cataract surgery, excluding eyes with cataracts and associated conditions that might predispose the development of glaucoma.

The bimodal onset of glaucoma after cataract surgery in this study (Fig. 1) reaffirms the observations of Mills and Robb,7 who also showed a bimodal pattern of onset of glaucoma following cataract surgery in children. They noted that glaucoma with an angle-closure mechanism frequently occurs within the first few months after surgery and open-angle glaucoma has a later onset (average 7.4 years). In contrast, Chen et al11 did not observe a bimodal pattern of onset of glaucoma and filtration angles were open in 94% of their patients. In our study group, 54% of eyes with early-onset glaucoma and 15% of the eyes with delayed-onset glaucoma had angle closure.

Simon et al5 reported that open angle glaucoma occurs after a latent period of 5.5 years following cataract surgery in children. Asrani and Wilensky9 reported an average interval between cataract surgery and the onset of open angle glaucoma of 12.2 years. In this study, with a mean follow-up period of 15.3 years, open-angle glaucoma developed between 6 and 14 years after cataract surgery. The time to the onset of open angle glaucoma was longer when patients were followed for a longer period. Children who have undergone cataract surgery have a continuing risk of developing glaucoma throughout their lives.2,5,8,9,11,13,17,18

Cataract surgery within the first year of life has been identified as a risk factor for glaucoma.3,7,10,11,12 This was also obvious in our aphakic glaucoma series; 92% of cataracts were removed before 1 year and 27% were removed before 1 month of age. Rabiah13 found that the incidence of glaucoma was 37% if the cataract surgery was performed before 9 months, 14% if the surgery was performed between 9 months and 2 years of age, and only 9% if performed between 2 to 3 years of age. Vishwanath et al12 found that after bilateral lensectomy within the first month of life, the 5 year risk of developing glaucoma in at least one eye was 50% but it decreased to 14.9% if the surgery is performed later in the first year. In this study, however, there was no statistically significant relationship between age at the time of cataract surgery and the type of glaucoma onset. More cases are needed to more fully study this correlation.

The timing of cataract surgery in infants may need to be reassessed. Prompt removal of congenital cataracts is considered to be important to achieve good visual and functional results19,20 although early surgery increases the risk for glaucoma. Vishwanath et al12 have suggested delaying surgery until the infant is 4 weeks old in bilateral cases, and we concur on the basis of these findings. Mori et al21 reported 58 eyes in 41 patients with congenital cataracts that were not surgically removed. During a mean follow-up of 19 years, only 2 eyes developed glaucoma. In both cases, glaucoma was of the angle closure type related to PHPV. There are three reports of patients with bilateral congenital cataracts, but only one of the eyes was operated on.2,5,7 In these patients, glaucoma developed only in the surgically treated eyes. Thus, glaucoma in children with cataracts seems clearly to be related to cataract surgery.

Early-onset glaucoma is strongly associated with the surgical procedure used to remove the cataract. In this study, pupil block with secondary angle closure was the principal cause of early-onset glaucoma. Mills and Robb7 made similar observations. Residual lens material contributes to the development of glaucoma by forming Elschnig pearls that mechanically block the pupil and induce inflammatory adhesions in the angle and at the pupil edge which can cause papillary block and angle closure.2,8,14,15 In this study, residual proliferating lens material needed to be removed later in 8 eyes. The conversion from assisted aspiration to outcome-enhanced aspiration combined with posterior capsulectomy and anterior vitrectomy has enabled the surgeon to remove the lens cortex and capsule more effectively. This change has reduced, but not eliminated, the prevalence of pupillary block and subsequent angle closure. In eyes that have poor pupil dilatation or when technical problems occur that preclude complete removal of the cortex, we recommend performing an iridectomy.

We found that sulcus-fixation of the IOL is a risk factor for the development of pupillary block, which has already been observed.7,22,23 In 3 of the 6 eyes in which the IOL was fixed in the sulcus, pseudophakic pupillary block developed. We recommend placement of the IOL within the capsular bag whenever possible. If this positioning cannot be accomplished, we recommend performing a peripheral iridectomy.

Microcornea is one of the risk factors for glaucoma that cannot be altered. An association between microcornea and development of glaucoma has been reported.7,9,10,12,13,16 Parks et al6 found a 33% incidence of glaucoma in eyes operated on that had a corneal diameter less than 10 mm. The prevalence of microcornea among aphakic glaucoma patients is as high as 88.5%6 to 94%.16 In our population, age adjusted microcornea was present in only 7% of eyes with early-onset glaucoma, but in 50% of eyes with delayed-onset glaucoma. Smaller eyes and eyes with reduced corneal diameter have a predisposition for angle closure. However, no eye with microcornea developed angle-closure glaucoma in our population; all had open angle glaucoma. Microcornea may account for the abnormally thick corneas seen in 11 eyes in this study. The increased thickness may cause the measured IOP to seem higher than it actually is. Although many of the eyes in our study with glaucoma had a thick cornea, we felt that the changes observed in the optic nerve cup, visual field changes, and occasional spikes in IOP over 40 mmHg meant that these eyes had glaucoma but not ocular hypertension.

Mori et al12 suggested that the aphakic state of the eye that follows the operation may be responsible for glaucoma. The volume of the lens and its dynamic role in accommodation are more prominent in younger eyes. Surgical removal of the lens early in life can alter normal development of the filtration angle. Morphologic studies of developing eyes of children have shown that the angle recess of the iridocorneal angle is expected to move toward the periphery, exposing the scleral spur and the ciliary body band.24,25 This study, Walton,8 and Keech et al4 show that there is altered development of the filtration angle. Eyes with delayed-onset open-angle glaucoma had an open angle, but there was anterior insertion of the iris together with the lack of development of the angle recess and filtration structures compared to adults. In the enucleated eye, we observed anterior displacement of iris tissue with incomplete development of the angle recess (Fig. 3). The trabecular meshwork was compact and thickened. These findings, and the changes observed on gonioscopy, suggest that there is a relative arrest in the normal development of the filtration angle and the trabecular meshwork in the eyes of aphakic children.

Alterations of the accommodation mechanism related to the loss of lens may also have a role in the functioning of the filtration angle. During accommodation, the longitudinal fibers of the ciliary muscle, which insert on the scleral spur, contract. This expands Schlemm's canal and opens the interstices of the trabecular meshwork, which increases aqueous outflow.26-30 Parasympathomimetic drugs stimulate the ciliary muscle, putting traction on the trabecular sheets which causes them to separate. These agents stimulate accommodation and are effective in lowering the IOP in aphakic eyes.

In conclusion, early-onset glaucoma after cataract surgery in children is usually of the angle-closure type secondary to pupillary block. Preventable risk factors for early-onset glaucoma include early cataract surgery, not performing an iridectomy when there is retained lens material after cataract extraction, a pupil that dilates poorly, a shallow anterior chamber, and sulcus fixation of an IOL. Delayed-onset glaucoma generally occurs years after surgery and is usually of the open angle type. Because of the late onset of this type of glaucoma, it is probably not directly linked to the surgery itself. Eyes with delayed-onset glaucoma have filtration angles that appear to be open yet also have findings consistent with incomplete development of filtration structures. Cataract surgery done at an early age and microcornea are risk factors for delayed-onset glaucoma.