Aim: To analyse cases of recurrent ectasia in donor corneas after penetrating keratoplasty (PK) for keratoconus.
Methods: Data on 25 patients (36 eyes) with recurrent ectasia were retrospectively analysed in this study. The main outcome measures were time to development of recurrent ectasia after first PK for keratoconus, change in keratometric sphere and astigmatism between final suture removal and development of recurrent ectasia, status of regrafts for recurrent ectasia, and histopathology of grafts excised for recurrent ectasia.
Results: The age at first PK was 32.6 (SD 8.5) years, and ectasia developed 21.9 (7.0) years after PK. The mean keratometric sphere and cylinder increased by 4.2 D and 3.0 D, respectively, between final suture removal and diagnosis of recurrent ectasia. Ectasia was often preceded by thinning without bulging of the recipient stroma at the graft–host junction. Fifteen eyes (13 patients) were regrafted for recurrent ectasia, and histopathology of the excised grafts showed changes characteristic of keratoconus in the donor tissue in all cases. Two regrafts (two eyes of one patient) developed ectasia again, with one eye requiring a third PK to improve vision.
Conclusions: Recurrent ectasia was diagnosed on average two decades after PK. Ectatic changes were often bilateral and occasionally recurred after regrafting, suggesting that host cellular and/or biochemical factors may be responsible. Repeat PK for recurrent ectasia is successful in the intermediate term.
Statistics from Altmetric.com
Keratoconus is frequently managed by penetrating keratoplasty (PK) when it is in its advanced stages.1 Late recurrence of keratoconus in the donor cornea is a known phenomenon, but published reports thus far have consisted of only small case series, as this disorder is infrequent.2–7
We retrospectively studied the records of all patients in our university-based practice with recurrent ectasia in PKs for keratoconus, highlighting the clinical features, histopathological findings, and the outcomes of repeat PK.
A series of patients with recurrent ectasia in corneal grafts performed for keratoconus was gathered, as they returned for follow-up examinations. The Corneal Transplantation Registry of the University of Michigan WK Kellogg Eye Center and the Department of Pathology databases were also searched for repeat corneal grafting in patients who had undergone prior PK for keratoconus. Institutional Review Board approval was obtained for retrospective review of medical data.
Data from patient records were analysed for age at first surgery, time to diagnosis of recurrent ectasia, gender and laterality. Best-corrected visual acuity (BCVA), spectacle-corrected if not routinely wearing a contact lens, as well as refractive and keratometric sphere and astigmatism were recorded.
The original diagnosis of keratoconus was confirmed on the basis of typical clinical signs, including inferior corneal thinning with cone, Vogt striae or Fleischer ring, and ancillary findings on keratometry and topography, if available. The diagnosis of recurrent ectasia after PK was based on clinical findings, such as inferior paracentral corneal thinning, Vogt striae, visible anterior bulge, irregular astigmatism, and computerised topographic findings, as well as by histopathological findings in the excised donor corneal button. Patients with high regular astigmatism on topography and those with inferior thinning restricted to only the recipient side of the graft–host junction were not included. No patients had a history of trauma to the graft that may have resulted in wound dehiscence.
For patients who underwent repeat PK for recurrent ectasia, the excised grafts were cut orthogonal to the cornea immediately adjacent to the region demonstrating maximal ectasia on gross examination. After processing the specimens routinely, 6 μm paraffin step sections taken at 50 μm intervals through the region of ectasia were stained with haematoxylin and eosin (H&E) and periodic acid–Schiff (PAS). The histopathological slides were examined for features of keratoconus. The clinical outcomes of these patients were also reviewed after repeat corneal grafting.
Twenty-five patients (36 eyes) with PK for keratoconus were identified with recurrent ectasia in the corneal graft (table 1). Fourteen patients (19 eyes) had PK performed elsewhere, but postoperative care was eventually transferred to our institution 7 to 32 years after surgery. The complete ophthalmological records were not available for these patients. Thirteen patients were male, and 20 had bilateral PK. The mean age at first diagnosis of keratoconus was 20.9 (SD 3.6) years (range 15–29), and age at first PK was 32.6 (8.5) years (range 15–53 years).
The mean recipient graft diameter was 8.1 (0.8) mm (range 6.0 to 9.0 mm), and the mean donor diameter was 8.2 (0.9) mm (range 6.0 to 9.0 mm). Suturing techniques included double-running (continuous 10-0 nylon with a second continuous 11-0 nylon) sutures in seven eyes, all interrupted sutures in seven eyes, a combination of interrupted and running sutures in two eyes, single running suture in one eye and unknown technique in the 19 eyes operated elsewhere.
Suture removal was typically complete by 1 to 2 years after surgery. The average keratometric sphere and cylinder shortly after all sutures were removed were 44.0 (1.9) D and 5.0 (3.5) D, respectively (n = 17), and the median best-corrected visual acuity (BCVA) was 20/25 (range 20/15 to 20/80) (n = 17).
Recurrent ectasia in the corneal graft was diagnosed 21.9 (7.0) years after PK (range 6 to 36 years). The average keratometric sphere and cylinder at the time when recurrent ectasia was diagnosed were 48.2 (5.0) D and 9.0 (4.4) D, respectively (n = 36). Ectasia was often preceded by thinning (without bulging) of the recipient corneal stroma at the graft–host junction, and the most frequent location was inferiorly (n = 20, table 2).
Fifteen eyes of 13 patients required a second PK to improve poor vision or contact lens intolerance caused by recurrent ectasia (table 3). The median BCVA just before regraft was 20/80 (range 20/20 to 20/400). The average keratometric sphere and cylinder just before regraft were 50.0 (5.9) D and 10.5 (4.1) D, respectively, and at last follow-up 45.9 (4.6) D and 5.0 (2.4) D, respectively (p<0.05; fig 1). The mean donor regraft diameter was 8.9 (0.3) mm (range 8.5–9.0 mm), which was a mean increase of 1.0 mm in button diameter over that of the previous graft.
Histopathology of all 15 specimens excised at repeat PK demonstrated features of keratoconus in the donor tissue. In 12 specimens, typical histopathological features of keratoconus were present, while atypical histopathological features were present in three specimens according to the criteria reported by Scroggs and Proia.8 All of the specimens demonstrated stromal thinning and scarring in regions of ectasia. In 12 specimens, typical alterations due to keratoconus included one or more breaks in the Bowman layer associated with irregular kinking and overlying epithelial thinning. The anterior stromal lamellae demonstrated Z-shaped irregularities in most of these specimens. In three specimens, atypical features consisted of ectasia accompanied by marked thinning of the Bowman layer with lesser degrees of epithelial thinning. In all 12 specimens, anterior stromal scarring extended through breaks in the Bowman layer (fig 2) in typical cases or replaced regions of the thinned Bowman layer in atypical cases.
Twelve of the 15 regrafts (11 patients) were doing well at last follow-up (6.9 (6.0) years, range, 6 months to 19 years). The median best-corrected visual acuity at last follow-up was 20/40 (range 20/20 to 20/100) with sutures remaining in three eyes. One regraft failed after 7 years due to graft rejection, and two regrafts (both eyes of one patient) developed ectasia again 19 and 21 years after being regrafted. This patient required a third PK in one eye to improve vision.
Patient 3 was diagnosed as having keratoconus at age 24 years in 1973. Her medical history was significant for asthma. She wore hard contact lenses until she required a PK in the right eye in 1976 (8.0 mm donor, 8.0 mm recipient, single-running suture). Preoperative keratometry was >60 D. Keratometry after the running suture was removed in 1978 was 41.75/50.50 @ 122°. Progressive increase in keratometric astigmatism was noted from 8.75 D in 1978 to 15.0 D in 2000 (fig 3), when keratometry was 42.00/57.00 @ 148°. Inferior ectasia was visible in the graft at this time but was not present on earlier routine postoperative examinations (fig 4). Ectasia was preceded by thinning without bulging of the host cornea at the inferior graft–host junction in 1992. The astigmatism continued to increase, and in 2003 the patient became intolerant of her contact lens and required a second PK (8.75 mm donor, 8.5 mm recipient, combination of interrupted and running sutures) to improve her reduced BCVA of 20/60. Histopathology of the excised graft showed significant corneal stromal thinning with irregularity and focal interruptions in the Bowman layer in both donor and host (fig 2A). At last follow-up in the regrafted eye, BCVA was 20/30, and keratometry was 46.50/48.00 @ 011° with sutures remaining.
PK (9.0 mm donor, 8.5 mm recipient, double running sutures) was performed in the contralateral (left) eye in 1984. Preoperative keratometry was >60 D. Keratometry after both sutures were removed in 1985 was 41.25/47.25 @ 098°. By 2003, keratometric astigmatism decreased in this eye, but the overall steepness by keratometry increased to 46.25/48.00 @ 034°, and inferior graft ectasia was apparent. However, the BCVA in this eye at last follow-up was 20/20, and the patient tolerated her contact lens well.
Patient 7 had a PK for keratoconus in each eye at age 33 years. Surgery was performed at another institution, and complete records were not available. Recipient trephine size was 6.5 mm in the right eye and 6.0 mm in the left eye. Eighteen years after surgery, the patient became intolerant of the contact lens in the right eye. The visual acuity was 20/25 with the contact lens, keratometry was 44.50/57.00 @ 160°, and visible ectasia was present at the inferior graft–host junction.
A second PK was performed in the right eye at that time (9.0 mm donor, 8.5 mm recipient, double running sutures). Keratometry after both sutures were removed a year later was 45.00/51.75 @ 166° but progressed to 41.75/55.00 @ 008° by 10 years after the second PK. Astigmatic keratotomy was performed twice, at 10 and 11 years after the second PK, after which keratometry improved to 48.00/49.00 @ 090°. Subsequently, astigmatism and keratometry increased again, and 19 years after the second PK, keratometry was 53.50/60.00 @ 076°, with visible ectasia in the graft.
A third PK was performed in the right eye at that time (9.0 mm donor, 8.5 mm recipient, interrupted sutures). Histopathology of all excised grafts showed irregular and focal interruptions of the Bowman membrane (fig 2B). Keratometry after all sutures were removed a year after the third PK was 39.75/43.25 @ 175°. The patient was seen in follow-up for 2 years after the third PK with little subsequent change, and BCVA was 20/40, until the patient died a short time later.
The left eye was noted to have ectasia in the graft at 19 years after the first PK, at which time keratometry was 38.25/56.50 @ 172°. A second PK was performed in that eye 20 years after the first PK (9.0 mm donor, 8.5 mm recipient, double running sutures). Keratometry after both sutures were removed 2 years after the second PK was 44.00/44.50 @ 002°, but this progressed to 36.0/44.5 @ 020 at last follow-up 19 years after the second PK, when inferonasal ectasia was noted, and BCVA was 20/50.
Data from this large series of patients with recurrent ectasia in penetrating grafts for keratoconus indicate that recurrence is often bilateral and typically occurs two decades after PK. Unfortunately, we were unable to estimate the incidence of recurrent ectasia in our population, because of inadequate data on the number of grafts done for keratoconus and likely bias in patient referrals. Recurrent ectasia in penetrating grafts has been previously reported as a rare occurrence,2–7 although a recent report suggested the incidence was as high as 11%.9 In many of these cases of “recurrent keratoconus,” histopathological features of keratoconus were seen in excised corneal grafts, and ectasia became evident approximately 20 years after PK, similar to our study. Table 4 compares the findings in our study to two other small series with clinical and histological descriptions of recurrent ectasia.
In our series, six grafts (five patients) had a recipient diameter ⩽7.0 mm, suggesting that inadequate excision of the cone with a small graft may have contributed to the development of recurrent ectasia (Case 7). Nevertheless, histopathology of the excised grafts demonstrated features consistent with keratoconus, suggesting that inadequate excision of abnormal tissue was not the sole cause of recurrent ectasia. Several cases were initially characterised by years of progressively increasing regular astigmatism. The astigmatism was sometimes mirrored about the vertical midline in both grafted eyes, suggesting recurrence of the disease rather than simple wound slippage. Progression of regular astigmatism in penetrating grafts for keratoconus has been reported with onset beyond the first decade after surgery, and may eventually result in recurrent ectasia with continued follow-up.11–13 Notably, in a recent report of PKs performed by a single surgeon with the same suturing technique, delayed progressive astigmatism was noted after PK for keratoconus but not after PK for Fuchs dystrophy, indicating that surgical technique and wound slippage are unlikely to contribute to the progressive keratometric changes.12 Our series is also notable for the number of bilateral cases of recurrent ectasia, and also for multiple recurrences of ectasia (Case 7).
Although 12 of the 15 repeat grafts in our series were surviving at last follow-up, the mean follow-up for these eyes after repeat grafting was only 5.0 (4.4) years, with the longest follow-up time of 15 years. Ten of the eyes had ⩾20/40 visual acuity, better than that reported in another study after repeat grafting for keratoconus.14 With longer follow-up, some of the second grafts are likely to become ectatic again, and a third PK may be required. Keratoconus has a low rate of overall graft failure compared with other indications for PK, but because keratoconus patients are younger than patients with other indications, the likelihood of needing more than one penetrating graft per eye in a lifetime is significant.15 It may be prudent, therefore, to defer PK for as long as possible in young keratoconus patients, although very-long-term data are not available to assess this approach. Nevertheless, intolerance of corneal contact lenses in primary or recurrent ectasia can be successfully managed with larger scleral lenses16 17 or with contact lenses with a special back surface design18 before considering surgical treatment.
Similar to the aetiology of primary keratoconus, the aetiology of “recurrent keratoconus” is unknown, but theories include inadequate initial excision of keratoconic tissue, grafting donor corneas that have forme-fruste keratoconus, or progression of the original disease process.19 20 Because recurrences in patients were often bilateral and multiple recurrences in the same patient were observed, the aetiology is likely to be related to host factors. Recurrent ectasia did not manifest clinically for two decades in many of the patients, which is similar to the time required for virgin corneas to develop clinical keratoconus. Keratoconus in virgin corneas probably results from a combination of keratocyte biochemical dysfunction,21 22 and keratocyte loss23 24 through apoptotic mechanisms.25 After PK, there is evidence that host keratocytes replace donor keratocytes26 and that keratocyte density in donor corneas is lower than in normal corneas.27–29 We postulate that over the two decades after PK for keratoconus, donor keratocytes are replaced by a low density of abnormal host keratocytes, and as a result, both the host and donor stroma become thinner at the weak graft–host junction, contributing to recurrent ectasia. This hypothesis is supported by the observation that host thinning occurs 14 years after PK and precedes recurrent ectasia.30 Furthermore, in most cases in our series, thinning of the graft–host junction occurred inferiorly, where the cornea is known to be biomechanically weak.31 It is possible that the factors associated with recurrent ectasia after PK for keratoconus might be akin to those responsible for ectasia after keratorefractive surgery.32 It is likely that the host epithelium, which eventually covers the donor cornea, may contribute to physiological and histological changes in the Bowman layer of the graft. Indeed, all of the specimens removed from patients with recurrent ectasia in our study demonstrated histopathological findings identical to those seen when keratoconus occurs primarily, although atypical features were present in three of our specimens as reported in other studies.3 4 8 Nevertheless, disruption of Bowman’s layer in excised grafts is not specific for recurrent ectasia, and can be detected in grafts for keratoconus that fail prior to recurrent ectasia becoming clinically evident.10
If recurrent ectasia were caused by transplantation of donor corneas with forme-fruste keratoconus, some of the mate donor corneas of ectatic grafts would be expected to develop recurrent ectasia also. Unfortunately, we were unable to trace the outcome of eyes that received the mate tissue for the ectatic grafts. Similarly, ectasia would be expected to occur more frequently in grafts done for non-ectatic indications, but this was not the case in a recent large comparative series.12 Several patients in our series developed recurrent disease in both eyes after bilateral PK. The chances that these patients received donor corneas with forme-fruste keratoconus in both eyes is exceedingly small, making it highly likely that recurrence is due to host factors. The estimated incidence of recurrent keratoconus9 is also probably higher than the incidence of forme-fruste keratoconus in the general population. Other factors contributing to recurrent ectasia might include biomechanical factors associated with asymmetric keratoconus, such as eye rubbing33 34 and side of sleep preference,35 36 and young donor age, which has been associated with steep keratometry.37 We were not able to retrospectively obtain the donor ages in 29 of the 36 grafted eyes, but the other seven donors were older than 31 years of age.
Recurrent ectasia was located inferiorly in most eyes in our series (fig 5). However, one eye developed recurrence superonasally but had histopathological changes consistent with keratoconus in the excised graft, and the fellow eye developed recurrence inferiorly. Therefore, although recurrent ectasia usually occurs inferiorly, it may also occur at other meridians, similar to primary, natural keratoconus.38–40 Although the clinical signs of recurrent ectasia may seem similar to those of graft wound dehiscence, histopathological features of keratoconus within the previous donor tissue argue against wound slippage, and none of our patients had a history of trauma or acute change in vision consistent with wound dehiscence.
In conclusion, recurrent ectasia was diagnosed on average two decades after PK, was often bilateral and occasionally recurred after regrafting; however, repeat PK for recurrent ectasia is successful in the intermediate term.
Funding: Supported in part by grants from Fight for Sight (GA07006), Midwest Eye-Banks, Research to Prevent Blindness (SVP as Olga Keith Wiess Special Scholar), and Michigan Institute for Clinical and Health Research New Pilot Clinical Trial Training Grant.
Competing interests: None.
Ethics approval: Ethics approval was provided by University of Michigan Institutional Review Board.
Patient consent: Obtained.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.