eLetters to:

ELECTRONIC ARTICLE: Naveed Hussain, Jonathan Clive, and Vineet Bhandari Current Incidence of Retinopathy of Prematurity, 1989-1997 Pediatrics 1999; 104: e26 [Abstract] [Full text] [PDF]

eLetters: Submit a response to this article

eLetters published:

Methods questions to clarify results

Dale L Phelps   (13 September 1999)

Response to Dr. Phelps'' comments.

Naveed Hussain   (5 October 1999)

Possible influences on ROP-incidence figures.

Jacqueline Termote, "Nicoline Schalij-Delfos, Bernard P Cats"   (28 November 1999)

Risk factors for severe ROP: Response to Drs. Termote, Schalij-Delfos and Cats

Naveed Hussain   (24 December 1999)

Methods questions to clarify results 13 September 1999
Dale L Phelps, Neonatology University of Rochester Send letter to journal: Re: Methods questions to clarify results dale_phelps{at}urmc.rochester.edu Dale L Phelps	To the Editor: The carefully collected and thoroughly presented epidemiologic data on ROP from a large tertiary center in Connecticut is a welcome addition of information for those caring for premature infants, and for those who attempt to develop rational screening criteria. I enjoyed working my way through the several interesting presentations of these data, and in the process identified some things that must be clarified. 1) In the Eye examination schedule, there are three criteria named for identifying infants who are to be examined. Does meeting any one criteria cause the infant to be examined, or does it require all three? This is particularly important with the first two criteria. 2) In the descriptions of threshold ROP given in the methods, the authors fail to mention the requirement of "plus disease" in order to make the diagnosis of "threshold ROP", although this was a requirement in the CROP-ROP study. This leaves the reader wondering if the authors actually failed to require plus disease for that diagnosis, or if this is an inadvertent omission from the description in the methods. 3) In Table 1, describing the baseline characteristics of those infants with ROP compared to those without ROP, several of the percentages are not what one would normally expect in such a table. They are, however, accurately described so that I could recalculate what I am more used to, but other readers may be confused. For example, in reporting on race, the table shows that 14.9% of infants with ROP were black(30/202), while 13% of infants without ROP were black(97/748). More conventionally, one would report that 30/127 (23.6%) of black infants developed ROP, while 172/823 (20.9%) of non-black infants developed ROP. The conclusions are similar, but the direction of the thought process is different. 4) In reference to the question of ROP and race, the CRYO-ROP study reported no difference in the proportion of black (63.2%) vs white (67.3%) infants who developed any stage of ROP, but rather observed that black infants progressed to threshold ROP less often than white infants (3.2% vs 7.5% of all infants, or 5.0% vs 11.1% of ROP cases).(1) I count 10 total infants who were treated with peripheral ablation for threshold ROP in this report, and this is likely too few a number to be analyzed for an effect of race. 5) In Figure 3A, the numbers at the top indicate "the sample size", but are a puzzle. These numbers, particularly in the higher gestational ages are far too small to represent all survivors (at less than 28 weeks they are almost identical). Comparing them to Figure 3B, they are not the number who developed ROP. They are probably the number of infants at each gestation who met examination criteria and were examined, since the text does say that "....the graph shows the relationship between GA and severity of ROP (percent of infants studied)." However, if one takes the number of cases of ROP for each gestation and divides by the "sample size", you do not get the percentages shown. You do however, get the percentages shown if you take the number of cases of ROP and divide by the number of survivors (calculated from figure 2), consistent with the numbers reported in the text describing the curve. The legend should have clarified this for the reader. Importantly: since many infants of gestational ages greater than 28 weeks were never examined, we must all remember use the incidence rates of ROP among these older infants with some caution. Sincerely yours, Dale L. Phelps, MD Reference List (1) Palmer EA, Flynn JT, Hardy RJ, Phelps DL, Phillips CL, Schaffer DB et al. Incidence and early course of retinopathy of prematurity. The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Ophthalmology 1991; 98(11):1628-1640.
Response to Dr. Phelps'' comments. 5 October 1999
Naveed Hussain, Neonatologist University of Connecticut School of Medicine Send letter to journal: Re: Response to Dr. Phelps'' comments. hussain{at}nso1.uchc.edu Naveed Hussain	October 5, 1999. We would like to thank Dr. Phelps for her thoughtful comments on our manuscript. She has brought to discussion some important issues that need clarification. 1. Clarification of criteria for eye exams. The eye exams were done for all infants who met the following criteria: (1) All infants < 30 weeks gestational age or < 1300 grams at birth; (2) All infants < 35 weeks gestational age or < 1800 grams at birth who received supplemental oxygen for more than a week. (3) Any newborn who remained in supplemental oxygen for > 60 days. The infant had to meet any one of the three criteria to be eligible for screening for ROP. 2. Was 'Plus disease' a requirement for 'threshold ROP'? No. Not all infants with 'threshold ROP' had 'Plus Disease'. Although the presence of 'Plus disease' was considered a marker for significant ROP. 3. .A different way of looking at the data on incidence of ROP in black infants. We agree that another way to describe the data would be to show that 30/127 (23.6%) of black infants developed ROP while 172/823 (20.9%) of non -blacks developed ROP. The conclusions are similar. However, we were specifically asked by one of the reviewers to clarify the incidence in black vs white infants and therefore we avoided the terms non-blacks or non-whites and reported specific data for whites and blacks separately. 4. Effect of race on need for ablation surgery We agree that there were not enough infants who needed retinal ablation surgery to analyze the effect of race in our report. We therefore reported only the overall incidence of ROP and noted that our findings were different than those of Saunders et al.1 5. Clarification of Figure 3. In figure 3A the numbers shown at the top are the number of infants who were examined and the bar graph illustrates the percent of infants with different grades of ROP. On the other hand, as is mentioned in the legend, in Figure 3B the absolute numbers of infants are shown in each bar graph. For example at < 23 wk GA n=21, with 71.4% i.e. 15 of them developing ROP (fig 3A); subsequently, Figure 3B then shows the number of these 15 infants that did or did not reach threshold ROP. We apologize for the lack of clarity of this description in the figure legend. 6. Caution in interpreting incidence in infants > 28 wk GA. Regarding the interpretation of ROP incidence in infants > 28 week GA, we examined 68%, 70%, 42%, and 29% of infants at 29-32 wk GA respectively who were the most at risk for ROP. The lack of need for retinal surgery in any infant in this population is certainly reassuring. However, there might be regional differences in ROP incidence and these need to be taken into consideration in interpreting these data. Reference: 1. Saunders RA, Donahue ML, Christmann LA, et al. Racial variations in retinopathy of prematurity. Arch Ophthalmol. 1996; 115:604-608.
Possible influences on ROP-incidence figures. 28 November 1999
Jacqueline Termote, Neonatologist, Pediatric Ophthalmologist and Neonatologist. University Medical Centre Utrecht, The Netherlands., "Nicoline Schalij-Delfos, Bernard P Cats" Send letter to journal: Re: Possible influences on ROP-incidence figures. B.P.Cats{at}wkz.azu.nl Jacqueline Termote, et al.	To the Editor, We read your article "Current incidence of Retinopathy of Prematurity, 1989-1997" with great interest. In a recent paper, accepted for publication in the Journal of Pediatric Ophthalmology and Strabismus, we reported about the incidence and severity of ROP in our tertiary care unit in Utrecht (The Netherlands), between 1986 and 1995. The study period was divided in two consecutive 5 year periods '86-'90 and '91-'95, to evaluate the effects of new methods of treatment (surfactant replacement therapy (SRT) and high frequency oscillatory ventilation (HFOV)) and general improvements in quality of care during the second period. In our study the overall incidence of ROP in the years 1986-1990 was not significantly different from the incidence of ROP in the years 1991-1995 (89/291 = 30.6% versus 172/466 = 36.9% respectively). As in your study, the incidence of severe ROP (ROP stage 3 or more) was significantly lower in the second period (15.7% versus 6.4%, p=0.015). Using logistic regression, we found that only SRT was associated with a decreased risk for severe ROP, the factors HFOV and general improvements in quality of care had no influence. We also compared the incidence of severe ROP in the two periods, for ROP patients with and without RDS. In patients with RDS the incidence of severe ROP decreased significantly during the second period (20.3% versus 6,4%, p=0,008). For patients not suffering from RDS however, the incidence of severe ROP remained unchanged. 1. Did you introduce HFOV or other new methods of mechanical ventilation in your unit during the study period and did you investigate the effect of these new methods of treatment on the incidence and severity of ROP in your population? 2. Did you also look for significant risk factors for severe ROP in your population? 3, Is it feasable that the decrease in severe ROP in your population is also the result of a decrease in severe ROP in RDS patients selectively, which would concur with our observations as mentioned above? Sincerely yours, Jacqueline Termote, Nicoline Schalij-Delfos and Bernard Cats.
Risk factors for severe ROP: Response to Drs. Termote, Schalij-Delfos and Cats 24 December 1999
Naveed Hussain, Neonatologist University of Connecticut School of Medicine Send letter to journal: Re: Risk factors for severe ROP: Response to Drs. Termote, Schalij-Delfos and Cats hussain{at}nso1.uchc.edu Naveed Hussain	December 21, 1999. We would like to thank Drs. Termote, Schalij-Delfos and Cats for sharing their findings on ROP from Utrecht, Netherlands, and for bringing to discussion some possible influences on ROP-incidence figures. The specific questions asked and the responses are as follows: 1. Did you introduce HFOV or other new methods of mechanical ventilation in your unit during the study period and did you investigate the efficacy of these new methods of treatment on the incidence and severity of ROP in your population? We introduced the use of HFOV in our NICU in 1992 but this modality of ventilation was initially used in the "rescue" mode for infants with severe lung disease or those with air-leaks. We started to use "elective" HFOV and SIMV (synchronized intermittent mandatory ventilation) in late 1995. However, we have not yet analyzed the data regarding the use of these modalities and ROP. 2. Did you also look for significant risk factors for severe ROP in your population? We had not studied the risk factors specifically for severe ROP (Stage 3 or more) in our original paper. However, in response to your question we re-analyzed our data and the results from univariate analysis of risk factors are shown below. Data represents mean + SD for continuous variables with Student's t-test * for significance; or % for nominal variables analyzed with Chi-square test#. TABLE ROP =< Stage 2 ROP >= Stage 3 Description (N= 906) (N= 44) p Value GA (wk) 29 + 3 25 + 2 < 0.0001* BW (g) 1262 + 487 745 + 169 < 0.0001 SGA (%) 12 5 NS# Whites (%) 73.5 75 NS# Blacks (%) 13.6 9.1 NS# Male gender (%) 52.5 59.1 NS# Vaginal delivery (%) 32 35 NS# RDS (%) 68.8 79.5 NS# Surfactant (%) 50 64 NS# Severe BPD (%) 36 64 0.0002# Ventilator (d) 12 + 17 42 + 20 < 0.0001* CPAP (d) 9 + 13 16 + 20 0.0006* Oxygen (d) 46 + 46 104 + 54 < 0.0001* NEC (%) 5.4 6.8 NS# Sepsis (%) 19.9 50 < 0.0001# IVH (%) 3.9 9.0 NS# PVL (%) 2.2 4.5 NS# Length of stay (d) 75 + 42 73 + 45 NS* 3. Is it feasible that the decrease in severe ROP in your population is also the result of a decrease in severe ROP in RDS patients selectively, which would concur with our observations as mentioned above? Unfortunately, we did not code for the severity of RDS and therefore are unable to evaluate the relationship between severe ROP and severity of RDS. Thank you for initiating this interchange of knowledge and ideas. Sincerely, Naveed Hussain, MBBS, DCH University Of Connecticut Health Center, Farmington, CT Jonathan Clive, PhD University Of Connecticut Health Center, Farmington, CT Vineet Bhandari, MD, DM Albert Einstein Medical Center, Philadelphia, PA