



* Neonatal Intensive Care Unit, Leicester Royal Infirmary, Leicester, United Kingdom
Department of Child Health, Leicester University Medical School, Leicester, United Kingdom
Department of Epidemiology and Public Health, Leicester University Medical School, Leicester, United Kingdom
|| Department of Ophthalmology, Leicester University Medical School, Leicester, United Kingdom
| ABSTRACT |
|---|
|
|
|---|
grade 3) retinopathy of prematurity (ROP) in infants with birth weight of
1250 g in a geographically defined population over a 10-year period.
Methods. An observational study was conducted of all infants who had a birth weight
1250 g and were born to mothers who were resident in the county of Leicestershire, United Kingdom, during the period January 1, 1990, to December 30, 1999. Cases were identified by the Trent Neonatal Survey. The incidence of severe ROP (
grade 3) was compared in 2 successive 5-year periods: 19901994 and 19951999.
Results. Comparing the first 5-year period (19901994) with the second (19951999), the total number of live births fell (60 789 vs 56 564). However, there was a significant increase in the number of births with birth weight
1250 g (including live and dead; 615 vs 734; live births only: 455 vs 556). Survival to 42 weeks of infants who were born at
1250 g was significantly better in the latter time period (203 vs 302; odds ratio [OR] for death: 0.54; 95% confidence interval [CI]: 0.390.75). The number of cases of severe ROP was 4 times higher during the second time period compared with the first (9 vs 36). A significantly increased risk for the development of severe ROP was seen during the second time period (OR: 2.92; 95% CI: 1.376.20). Even after allowing for the change in gestation induced by the improved survival during the second time period, the increased risk remained (OR: 2.81; 95% CI: 1.276.21).
Conclusions. There is strong evidence that the incidence of severe ROP among infants with birth weight
1250 g increased in the latter half of the last decade. The increased risk seems to be independent of the increase in survival.
Key Words: retinopathy of prematurity survival rate incidence
Abbreviations: ROP, retinopathy of prematurity TNS, Trent Neonatal Survey OR, odds ratio CI, confidence interval
Retinopathy of prematurity (ROP) is a disease that affects the formation of retinal blood vessels in premature infants and is a major cause of blindness.1 The incidence of this condition, in particular trends over time, has been the subject of much debate. The difficulties of consistent diagnosis and classification pose major problems in studies involving multiple observers, particularly those who try to identify all cases irrespective of grade. Of equal importance are issues of referral and inclusion bias in studies that are based on individual hospitals or groups of hospitals.27 Previous investigations by our own group have suggested that the incidence of severe ROP is related to the rate of survival of the most immature infants in a particular locality.8 However, although these findings were based on examinations performed by single observers in each of 5 different services, it was impossible to estimate numeric values for incidence without reliable denominators. We were keen to explore this issue further by examining trends in survival and rates of severe ROP in a geographically defined population over a 10-year period.
| METHODS |
|---|
|
|
|---|
1250 g and were born to mothers who were resident in the county of Leicestershire, United Kingdom, during the study period January 1, 1990, to December 31, 1999. Cases were identified by the Trent Neonatal Survey (TNS). This survey, established in 1990, is an ongoing study of neonatal intensive care activity in the Trent Health Region of the United Kingdom and has been described previously.9 All of the perinatal services in the Trent Region contribute to the TNS, and units in adjacent regions also permit data collection on Trent infants. Leicestershire County comprises approximately one sixth of the Trent Region with a population of
1 million and
10 000 births per year. The database holds information related to all infants who were of 1500 g birth weight or less and/or 32 weeks gestation or less and were born to a Trent resident mother and admitted to a neonatal unit since 1990. Data for the TNS are collected by a group of 5 part-time research nurses who visit each neonatal unit on a regular basis and complete a standardized data set about each infant. Information is obtained from the clinical records, discussion with staff, and, when appropriate, personal observation. Data from 2 additional ongoing surveys, the Leicestershire Perinatal Survey and the Confidential Enquiry Into Stillbirths and Deaths in Infants, provided data on infants who were born between 20 and 32 weeks gestation, were in the appropriate weight category, and were either 1) born dead or 2) live born but not admitted for intensive care having been considered to be nonviable. Data from each of these sources was cross-referenced with the admission books of the relevant neonatal units.
One of the authors (G.W.) was responsible for the eye examinations in Leicestershire during the whole period. Approximately 98% of the eligible cohort were examined by him personally or by his registrar under his supervision. The remaining infants from Leicestershire were cared for in hospitals in neighboring counties, and these infants were examined by similar specialist teams. Findings in all settings were recorded using the internationally recognized classification for ROP.10,11 For limiting the effect of interobserver and intraobserver error, inherent in any clinical scoring system, only infants who were identified as having ROP stage 3 and above according to the international classification of ROP were included in the study.
Eye Examination Schedule
All infants who were born with a birth weight of 1250 g or less underwent eye examinations. The first eye examinations were done when the infants reached a postmenstrual age of 32 weeks. Subsequent examinations were conducted at least every 2 weeks until at least 1 examination beyond 40 weeks of postmenstrual age. When stage 3 ROP was imminent or established, the infants were examined more frequently and for a more prolonged period. To ensure completion of the screening process, appropriate arrangements were made for infants who were due to be transferred to another hospital or discharged from the hospital.
Eye Examination Method
The pupils were dilated with cyclomydril eye drops (0.2% cyclopentolate in combination with phenylephrine 1%). The drops were instilled on 2 occasions at least 30 minutes before examination. Examinations were performed using indirect ophthalmoscopy with a 28-dioptre lens. An eyelid speculum and scleral indenter were used. Using the international classification of ROP the following were recorded: 1) ROP; 2) severity; 3) location by zone; 4) extent in clock hours; 5) "plus disease"; and 6) other clinical signs, including changes in cornea, anterior chamber, iris, pupil, lens, and vitreous. The findings were recorded in eye examination logbooks held by the ophthalmologist, and a copy of the record was also filed in the patient case notes.
Statistical Analysis
For the analysis, the maximum severity of ROP in any 1 eye for an individual infant was recorded. The study period of 10 years (19901999) was divided into 2 time periods of 5 years each, and the incidence of severe ROP among infants with birth weight
1250 g was calculated. The outcomes in the 2 periods 19901994 and 19951999 were compared using estimated odds ratios (ORs). The results were then studied in relation to the survival rates during these time periods. Potential confounding variables introduced into the models were gestation at birth, birth weight, and gender. Variable selection was assessed through changes in the log likelihood, and quadratic terms were introduced for the continuous variables to check for nonlinearity. For the purposes of this study, survival was defined as survival to 42 weeks postmenstrual age. All analyses were completed in SAS v8.00. The survival rates and incidence of severe ROP were calculated for the 2 time periods.
| RESULTS |
|---|
|
|
|---|
1250 g (both live and dead: 615 vs 734; P < .0001; Live births only: 455 vs 556; P < .0001). Of the live births with birth weight
1250 g, 321 and 397 infants were admitted for neonatal intensive care in 19901994 and 19951999, respectively. This was an identical proportion in both time periods. The remaining live-born infants were assessed by the neonatal team to be nonviable and remained with their parents. Of those infants who were admitted to the neonatal unit, 203 and 302 survived to 42 weeks postmenstrual age in the 2 5-year periods, respectively, resulting in a significantly improved survival in the latter time period (OR for death: 0.54; 95% confidence interval [CI]: 0.390.75; P = .0002).
|
grade 3 ROP, and among the 302 infants in the second time period (19951999), 36 infants were identified to have
grade 3 ROP. (OR for the development of severe ROP during the second time period: 2.92; 95% CI: 1.376.20; P = .0054). We were concerned that the improved survival might have resulted in an increased number of low birth weight, low-gestation infants (ie, high-risk infants) surviving to a stage at which they were eligible for eye examination (Figs 1 and 2). The OR for the risk of severe ROP hence was corrected to take into account this change in gestation and birth weight. However, the increased risk remained during the second 5-year period (OR: 2.81; 95% CI: 1.276.21; P = .011). Gender was not statistically significant at the 5% level (P = .27) and therefore was excluded from the final model, which contained terms for gestation (P = .005) and birth weight (P = .0009).
|
|
| DISCUSSION |
|---|
|
|
|---|
Although both the total number of infants who are offered intensive care and the proportion who survive have grown, neither of these factors seems to account for the observed change in the number of cases of severe ROP. It is important to note that the observed changes are based on a whole population and not a hospital or a group of hospitals, thereby eliminating referral and/or inclusion bias as a cause for this phenomenon. Although we have adjusted the observed rate of severe ROP for the increased survival, we have not attempted to take special account of the increased numbers of infants who survive at the very lowest gestations. The data related to the gestation of infants who developed severe ROP suggest that this would not be helpful.
There has been a great deal of discussion about the incidence of ROP after it was first described by Terry.12 Up until the early 1970s, apparent increases and decreases were thought to reflect changes to protocols governing the use of oxygen in at-risk infants.1316 During the 1970s and 1980s, the improved survival rate of extremely low birth weight infants associated with the advances in neonatal medicine was considered to be an important factor in the observed changes in incidence of ROP.17 Most studies at that time defined ROP in terms of the current incidence rather than change in incidence.1821 However, some studies did consider changes in incidence over time but were largely based on hospital populations. As a result, it would have been difficult to assess changes in the population characteristics of the type that we have demonstrated. Indeed, an inability to account for such a change may explain the contradictory nature of these studies with some showing an increased incidence3,5,6,21 and others a decrease.4,2229 The introduction of indirect ophthalmoscopy and routine postdischarge follow-ups were also potential confounding factors in studies that compared incidence in the 1960s and 1970s.3
More recent studies do not provide a clearer picture. Rowlands et al28 reported a significant decrease in the incidence of ROP. However, again, the work was based on a single center. A previous study conducted by our own group8 looked at the incidence of severe ROP in 5 tertiary care centers over 1 year. The study used severity of illness scoring in an attempt to "standardize" the populations. Although improved survival in 1 center seemed to correlate with an increased risk of ROP, there were concerns that the disease severity correction used in the study had shortcoming in relation to the most immature infants. Bullard et al23 compared their single-center data at Vanderbilt for the period in which they were 1 of 23 CRYO-ROP study centers (19861987) with the data from 19951996 and found that the incidence of all levels of ROP across all birth weights had decreased. The decrease in incidence and severity of ROP was attributed to the change in use of surfactant, continuous pulse oximetry, improved nutritional support, and use of antenatal steroids. Reynolds et al30 rejected this conclusion and suggested that the single-center findings from Vanderbilt University represented regression toward the mean. They also pointed out that the multicenter LIGHT-ROP trial31 conducted from 1995 to 1997 had shown no reduction in the incidence of ROP. Several studies, including randomized controlled trials,3239 have looked specifically at the effect of surfactant therapy on the incidence and or severity of ROP but have produced conflicting results.
Two studies in the 1980s prospectively investigated conventional incidence rates in geographically defined populations.40,41 The prospective population-based study by Darlow et al40 was conducted in New Zealand and assessed the incidence of ROP for all very low birth weight infants who were born in New Zealand over 1 year (1986). The second study was conducted prospectively in the time period July 1, 1985, to May 31, 1987, in the 5 neonatal units that serve the UK health authorities of Leicestershire, Nottinghamshire, and Southern Derbyshire.41 Given the major changes that have occurred subsequently in both the approach to neonatal care and survival of the most preterm infants, it is difficult to compare the result of these studies, meaningfully, with those that we have obtained. However, it is interesting that the rate of severe ROP in Ngs population (which included Leicestershire) was very low. Two other community-based studies of low birth weight survivors, from a similar era, were conducted in Canada.42,43 These were retrospective. Ophthalmology was only 1 aspect of these studies, which used a different classification system and hence could not be compared with other studies. More recently, a population-based Swedish study compared rates of ROP during 2 periods (19881990 and 19982000) and found an increased incidence of severe ROP. This study found that the incidence was particularly related to gestation but did not have data related to changes in the underlying population characteristics over time.44
The background to our findings, from the existing literature, therefore is somewhat inconsistent. We believe that we have demonstrated convincing evidence of an increase in the incidence of severe ROP during the 1990s. This change occurred against a background in which both the proportion of immature infants who were offered intensive care and the survival increased. The scale of the change in how the most immature infants were classified and treated has clearly been very large over this period. However, from our analyses, it seems that these factors alone are not sufficient to explain the change in the incidence of severe ROP. An alternative explanation is that the change represents regression of the figures for Leicestershire toward the mean. This seems unlikely because we have compared a range of neonatal outcomes for the Leicestershire population with other parts of the Trent Region annually since 1990 as part of the TNS. In addition, our previous work in relation to ROP suggested that Leicestershire was not atypical in relation to other parts of the United Kingdom.8 A final explanation is that some new factor has come into play, ie, that there is something different in current survivors that cannot be accounted for in terms of birth weight or gestation or that some aspect of treatment during the second study period led to an increased risk of severe ROP developing. Because of the nature of our study, we have no data on which to base such an etiologic explanation for the observed change. It is clear that markedly improved survival rates over the period make it unlikely that "improperly delivered neonatal care" is the reason for more immature infants developing severe ROP. In relation to specific changes and developments in neonatal care that have happened in the past 10 years, studies that have tried to explore such a link in relation to resuscitation,45 blood gas monitoring,46 and mode of delivery47 have not reached a clear answer.
Although there has been much debate about the changes in the incidence of ROP since it was first described, we believe that this study clearly shows an increase in the incidence of the most severe form of ROP in this population over the past 10 years. The pathophysiology of this condition is well understood.48 Clearly, the challenge now is to try to identify a mechanism for the observed change.
| FOOTNOTES |
|---|
Reprint requests to (D.J.F.) Neonatal Unit, Leicester Royal Infirmary, Infirmary Square, Leicester LE1 5WW, UK. E-mail: david.field{at}uhl-tr.nhs.uk
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
C. Slidsborg, H. B. Olesen, P. K. Jensen, H. Jensen, K. R. Nissen, G. Greisen, S. Rasmussen, H. C. Fledelius, and M. la Cour Treatment for Retinopathy of Prematurity in Denmark in a Ten-Year Period (1996 2005): Is the Incidence Increasing? Pediatrics, January 1, 2008; 121(1): 97 - 105. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. D. Modanlou, Z. Gharraee, J. Hasan, J. Waltzman, S. Nageotte, and K. D. A. Beharry Ontogeny of VEGF, IGF-I, and GH in Neonatal Rat Serum, Vitreous Fluid, and Retina from Birth to Weaning Invest. Ophthalmol. Vis. Sci., February 1, 2006; 47(2): 738 - 744. [Abstract] [Full Text] [PDF] |
||||
![]() |
Early Treatment for Retinopathy of Prematurity Coo The Incidence and Course of Retinopathy of Prematurity: Findings From the Early Treatment for Retinopathy of Prematurity Study Pediatrics, July 1, 2005; 116(1): 15 - 23. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||