Participants: Neonates with threshold retinopathy who underwent diode laser photocoagulation of the peripheral avascular retina.
Methods: A retrospective chart review was done of the weekly examination records of infants treated for threshold disease. Features that were studied included the presence of residual stage 3 neovascularization, plus disease, and development of retinal detachment (RD).
Main Outcome Measures: Timing of full involution and/or development of an RD.
Results: Of 262 eyes of 138 infants treated, full involution without RD was seen in 8%, 43%, 64%, 73%, and 86% of eyes at postoperative weeks 1, 2, 3, 4, and 9⫾3, respectively. Retinal detachments were diagnosed cumulatively in 0%, 1.5%, 4.2%, 6.5%, and 14% of eyes at weeks 1, 2, 3, 4, and 9⫾3, respectively.
Conclusions: Full involution of laser-treated threshold retinopathy of prematurity required more than 2
weeks in more than half of treated eyes. Most RDs were not detected until ⱖ3 weeks after
treatment. Ophthalmology 2004;111:1894 –1898 © 2004 by the American Academy of Ophthalmology.
Retinopathy of prematurity (ROP) is a retinovascular dis-ease that occurs exclusively in premature infants. Despite timely peripheral retinal ablation with cryotherapy or laser, a significant number of threshold eyes still progress to an unfavorable outcome.1Retinopathy of prematurity remains
a leading cause of severe visual impairment and blindness.2 A large body of research is available regarding important clinical and basic science aspects of ROP. Numerous risk factors have been identified, most prominent of which are low estimated gestational age and low birth weight.3 An international classification system is widely accepted,4,5and
screening criteria have been uniformly agreed upon.6 Despite being replete with information on the progres-sion and treatment of ROP, the literature on involution of ROP is limited. Repka et al7 and Preslan and Butler8
re-ported on regression of nonthreshold ROP. A literature search on PubMed failed to identify studies analyzing the process of involution after treatment of threshold ROP. Guidelines for observing treated infants are based primarily on experience. Determination of when the disease is re-gressing in a favorable manner and identification of unfa-vorable regression patterns are based primarily on experi-ence. Guidelines do not exist to advise clinicians on when an eye needs additional ablative treatment. Detailed infor-mation on the process of ROP involution after treatment of threshold disease could prove useful to clinicians in the management of neonates with ROP, and could impact de-velopment of reintervention strategies for eyes that are not involuting in a favorable manner. The purpose of this study was to analyze systematically and report the process of ROP involution after treatment of threshold disease with transpu-pillary diode laser photocoagulation.
Materials and Methods
A retrospective study of consecutive patients who underwent diode laser photocoagulation for threshold ROP at Texas Children’s Hospital and Texas Women’s Hospital between 1996 and 2000 was conducted. Eligible patients were identified through a logbook of laser-treated ROP patients. To be eligible for inclusion, the infants had to have received diode laser treatment for threshold ROP and had to have a birth weight ofⱕ1500 g or an estimated gestational age ofⱕ28 weeks, consistent with present screening recommendations.6
Follow-up for at least 9⫾3 weeks after
treat-Originally received: November 18, 2002.
Accepted: February 13, 2004. Manuscript no. 220916.
1Department of Ophthalmology, Cullen Eye Institute, Baylor College of
Medicine, Texas Children’s Hospital, Houston, Texas.
2Department of Pediatrics, Cullen Eye Institute, Baylor College of
Med-icine, Texas Children’s Hospital, Houston, Texas.
Supported by a grant from the Knights Templar Foundation, Chicago, Illinois, and an unrestricted grant from Research to Prevent Blindness, Inc., New York, New York
Presented in part at: American Academy of Ophthalmology Annual Meet-ing, November 11–14, 2001; New Orleans, Louisiana.
Correspondence to David K. Coats, MD, Baylor College of Medicine, Texas Children’s Hospital, 6621 Fannin MC CCC 640.00, Houston, TX 77030. E-mail: firstname.lastname@example.org.
ment and fundus examination at least every other week for the first 4 weeks after treatment were additional inclusion criteria. Eyes requiring retreatment for skip areas were excluded because of the desire to study the process of involution in fully treated eyes. Data analyzed included basic demographic information such as esti-mated gestational age and birth weight. Threshold ROP was con-firmed by identifying features consistent with threshold disease in the preoperative clinical diagram, clinical notes, and operative reports. The ROP treatment protocol during the years under study consisted of laser application to the entire peripheral avascular retina with light to medium intensity burns separated by no more than one spot size.
The status of the retinal examinations was tabulated and ana-lyzed for each treated eye during the first 4 weeks after surgery and again at 9 (⫾3) weeks after treatment, using the Kaplan–Meier estimator. Right and left eyes were analyzed separately. Because only a single examination was required after 28 days, events occurring after this time were interval censored, whereas events before this time were right censored. Although an examination 9 weeks after treatment might not be sufficient to detect all eyes that ultimately developed a retinal detachment (RD), we limited our data collection to these time limits because we were primarily interested in studying acute ROP involution and because patients were often dispersed from our tertiary care setting to continue follow-up at community locations near their homes after the acute disease process had resolved, making long-term disease analysis less reliable due to loss of patients to follow-up elsewhere.
Involution parameters were established before initiating data collection. To minimize investigator bias, evaluation of the post-operative retinal status was limited strictly to those features of active ROP that could be clearly and unmistakably identified from review of the clinical records. Complete involution of ROP was defined as the absence of any active neovascular tissues and absence of any residual dilation and/or tortuosity of the arterioles or venules in the posterior pole without development of an RD. Vitreous organization above the ridge with no obvious active vascular elements was not considered to be active neovasculariza-tion.
Active stage 3 neovascular disease was classified as present or absent. Because our standard clinical examination form relies on diagrams without numeric classification of each individual retinal sector, no attempt was made to count the number of clock hours with active stage 3 disease except at threshold. Residual plus disease was considered to be present if any dilation or tortuosity of the posterior pole vessels remained in any quadrant. Although limiting the classification of stage 3 disease and plus disease to a simple present or absent assessment does not reflect the full
dynamics of the involution process, it was felt that this approach was most appropriate for this retrospective study because it re-quired no subjective interpretation of the clinical documentation and thus minimized investigator bias. Eyes with vitreous hemor-rhage precluding visualization of the retina were considered to have active ROP until involution could be visualized or an RD was detected. Approval from our institutional review board was ob-tained.
A total of 262 eyes of 138 infants were analyzed. Sixteen eyes were excluded because retreatment for skip areas of untreated retina was required. The mean estimated gestational age of treated infants was 25.5 weeks (range, 22–36), and the mean birth weight was 760 g (range, 300 –1500). Cumulatively for both right and left eyes, ROP had fully involuted in 21 (8%) eyes 7 days after treatment (Fig 1). Cumulatively, complete involution without de-velopment of an RD was seen in 43%, 64%, 73%, and 86% of eyes at postoperative days 14, 21, 28, and 63⫾21, respectively. Stage 3 neovascularization generally involuted more slowly than plus dis-ease (Figs 2,3).
Retinal detachments were tabulated as present or absent. No attempt was made to segregate eyes based on severity of the detachment, because intervention by vitreoretinal surgery often occurred as soon as a detachment was noted, thus rendering evaluation of the natural history of detachment after treatment impossible. No RDs were diagnosed 1 week after treatment. De-tachments were seen in 1.5%, 4.2%, 6.5%, and 14% of eyes at 14, 21, 28, and 63⫾21 days, respectively (Fig 4). Of 16 eyes noted to have a dense vitreous hemorrhage at any examination during the first 4 weeks after surgery, 12 (75%) developed an RD.
Despite an effective treatment for ROP, the disease contin-ues to be an important cause of severe visual impairment and blindness.2Although the disease is often thought of as being defined by specific events such as prethreshold, threshold, and unfavorable or favorable outcome, it is not truly punctuated by these precisely defined events. Instead, a continuum exists throughout the disease process during both the progression and involution phase of the disease. Figure 1. Kaplan–Meier graph of full retinopathy of prematurity involution over time, with 95% confidence intervals.
Each phase is defined over a period of time in which the disease process evolves, with one phase blending seam-lessly into the next.
Useful and important guidelines are available for the practicing clinician regarding which children require sur-veillance for ROP. Appropriate tracking schedules based upon disease status and risk are well accepted.6Treatment criteria at threshold have been clearly established,1though new guidelines have recently been introduced.9Information is also available to help clinicians better understand how nonthreshold ROP involutes.7,8In contrast, we could find no detailed information regarding important aspects of disease involution after treatment of threshold ROP. How often should patients be examined after treatment? What is the normal rate and pattern of involution? What are the early signs that an unfavorable outcome will develop? When should retreatment be considered, and when should a vit-reoretinal surgeon be consulted? The answers to these im-portant questions are generally learned through experience and, to our knowledge, are not described clearly in the scientific literature.
No clear guidelines exist as to how frequently infants should be observed clinically after laser photocoagulation. Our own custom has been to observe infants weekly, when possible, until the disease has resolved or a detachment ensues. We found that the majority of RDs that occurred
were not detected until more than 3 weeks after treatment, a time after which active neovasculariation and active plus disease had already resolved in most cases. This has impor-tant implications for the management of treated ROP pa-tients and supports the contention that vitreous factors play a critical role in the development of cicatricial ROP, as suggested by Hikichi et al.10Vitreous traction seems to be able to exert considerable influence on the retina even after the acute ROP signs of neovascularization and plus disease have resolved.
Are clinically important signs predicting eventual RD detectable during the early involution process of ROP? A preliminary analysis of risk factor data from patients eval-uated in this study demonstrated that when the retina could not be visualized in any given week because of severe vitreous hemorrhage, the risk of ultimately being diagnosed with an RD was extremely high. Other risk factors may be important, and we are in the process of reviewing our data to report these findings. Functional outcome after treatment of cicatricial ROP is usually less than optimal.11,12 Given this unfortunate reality, identification of high-risk signs that predict the development of an unfavorable outcome after treatment of threshold ROP could potentially lead to the development of improved reintervention strategies. It is our hope that development of strategies that could be imple-mented preemptively to prevent RD before it occurs can be Figure 2. Kaplan–Meier graph of full involution of all elements of plus disease over time, with 95% confidence intervals.
developed, resulting in significant improvement of visual outcomes in high-risk eyes.
At present, no standard guidelines exist on how to doc-ument the involution process after treatment. We found, upon reviewing our data, that lack of a standardized ap-proach for documentation of regression factors limited our ability to derive important information from the review of our clinical records, thus limiting inferences that could be made from our data. We propose that at least 4 clinical factors should be noted at each examination after treatment, and propose simple methods to document these features:
1. Plus disease should be designated as present, resolv-ing, or resolved. The term present should be utilized when unequivocal dilation and tortuosity of arterioles and venules are present in at least 2 quadrants in the posterior pole. Resolved should be used to denote the absence of any dilation or tortuosity of any vessels in the posterior pole, and resolving to describe a vascu-lar status somewhere between these 2 extremes. Al-though this proposed classification scheme for plus disease does not convey the full dynamics of the involution process, it is practical in that it limits the number of subjective decisions required and provides at least limited information about the timing of invo-lution. Nothing short of photographic documentation is likely to result in information that is more useful or reliable.
2. Stage 3 disease should be documented per the Inter-national Classification of ROP,4 with careful docu-mentation of the number of clock hours in which neovascularization is still present. We recommend identifying neovascular tissue that has separated from the ridge separately. This classification scheme clearly outlines important involutional features of stage 3 neovascular disease, because reduction in the number of active clock hours and separation of the neovascular tissue from the ridge are both signs of disease involution.
3. The status of the vitreous adjacent to the vascular/ avascular junction should be noted. The features of significant vitreous organization are readily apparent
and are likely to play a key role in the development of an unfavorable outcome.10Much information can be easily imparted in simple clinical diagrams or descrip-tions. If no vitreous organization is seen, no specific documentation is required. The presence, number of clock hours included, and density of vitreous organi-zation can be represented in diagrammatic or descrip-tive formats. Although this level of documentation does not represent the full spectrum of vitreous orga-nization, it calls for very limited subjective interpre-tation by the examiner and, we feel, is not likely to be improved upon except with photographic documenta-tion, something that is not yet practical in most cen-ters.
4. Because severe vitreous hemorrhage seems to be an important prognostic sign, we suggest noting areas where visualization of the retina is prevented by vit-reous hemorrhage. This can be done in a simple annotated diagram or via a written description of the findings.
Only a long-term, prospective study can fully evaluate the process of involution and clearly identify favorable and unfavorable patterns of disease involution. In addition, a prospective study would allow better characterization of threshold disease and comparison of threshold severity with regression trends. Information regarding the preoperative severity of threshold ROP, such as the number of clock hours of stage 3 disease and number of quadrants of plus disease, could be useful when judging the process of invo-lution. Worsening of ROP, for example, might be indicative of the need for preemptive retreatment. It is not possible to provide a clear analysis of the preoperative threshold sever-ity in this study without subjective interpretation of the preoperative examination forms, and this type of data is best collected on a prospective basis.
The information derived from our study must be viewed in the context of several limitations. Because the study is retrospective, there were no strict guidelines for documen-tation during the period under study, though our group has many agreed-upon standards. This required us to limit in-terpretation of the data to those documented features that Figure 4. Kaplan–Meier graph of retinal detachment over time, with 95% confidence intervals.
on the process of involution of threshold ROP after transpu-pillary diode laser peripheral photocoagulation. Approxi-mately half of treated eyes involuted fully within 2 weeks of treatment, whereas the remaining eyes requiredⱖ3 weeks to involute fully. Dense vitreous hemorrhage totally obscur-ing visualization of the retina was a bad prognostic sign, eventually followed by diagnosis of RD in more than half of affected eyes. Retinal detachments were usually not de-tected untilⱖ3 weeks after treatment, and some were not detected until after complete involution of neovasculariza-tion and plus disease had occurred, supporting the noneovasculariza-tion that vitreous factors play an important role in the develop-ment of ROP detachdevelop-ments. Because RDs typically were not diagnosed in the immediate postoperative period, screening should be considered for a minimum of 10 weeks after treatment. Additional follow-up may be needed, depending on the severity of the disease. Further characterization of favorable and unfavorable patterns of ROP involution after treatment of threshold disease in a prospective study may provide a better rationale for scheduling follow-up exami-nations and could lead to the development of preemptive reintervention strategies designed to prevent RD from de-veloping in eyes with high-risk involution features. Such prophylactic measures are currently not recommended until more information is available to support this treatment approach.
rity. An international classification of retinopathy of prematu-rity. Arch Ophthalmol 1984;102:1130 – 4.
5. International Committee for the Classification of the Late Stages of Retinopathy of Prematurity. An international clas-sification of retinopathy of prematurity. II. The clasclas-sification of retinal detachment. Arch Ophthalmol 1987;105:906 –12. 6. American Academy of Pediatrics, Section on Ophthalmology.
Screening examination of premature infants for retinopathy of prematurity. Pediatrics 2001;108:809 –11.
7. Repka MX, Palmer EA, Tung B, Cryotherapy for Retinopathy of Prematurity Cooperative Group. Involution of retinopathy of prematurity. Arch Ophthalmol 2000;118:645–9.
8. Preslan MW, Butler J. Regression pattern in retinopathy of prematurity. J Pediatr Ophthalmol Strabismus 1994;31:172– 6. 9. Early Treatment for Retinopathy of Prematurity Cooperative Group. Revised indications for the treatment of retinopathy of prematurity: results of the Early Treatment for Retinopathy of Prematurity Randomized Trial. Arch Ophthalmol 2003;121: 1684 –94.
10. Hikichi T, Nomiyama G, Ikeda H, Yoshida A. Vitreous changes after treatment of retinopathy of prematurity. Jpn J Ophthalmol 1999;43:543–5.
11. Cherry TA, Lambert SR, Capone A Jr. Electroretinographic findings in stage 5 retinopathy of prematurity after retinal reattachment. Retina 1995;15:21– 4.
12. Trese MT, Droste PJ. Long-term postoperative results of a consecutive series of stages 4 and 5 retinopathy of prematu-rity. Ophthalmology 1998;105:992–7.