Retinopathy of prematurity (ROP) is a disease characterized by abnormal retinal vasculature that can have devastating visual consequences. Despite evidence that early detection and treatment can prevent blindness,1,2 ROP remains a leading cause of pediatric blindness worldwide.3–6 The “third epidemic” of ROP is characterized by higher overall survival rates of premature infants with a consequently higher incidence of ROP due to lack of appropriate neonatal intensive care unit resources needed to regulate inhaled oxygen delivery and prevent ocular complications.5,7,8 Digital imaging and the use of telemedicine is well-suited for screening and diagnostic purposes, especially in low- to middle-income countries where there are limited numbers of providers who can accurately diagnose disease requiring treatment.9
Previous studies have noted larger, older infants with ROP in low- and middle-income countries than would be expected in the United States and other high-income countries.8,10 Birth weight (BW) and gestational age (GA) have repeatedly been shown to be the most important risk factors for ROP and are the main criteria used in screening programs. However, it remains to be determined whether exact cut-offs should be specific to the population of interest or to determine criteria that are broadly applicable to lowand middle-income countries with a high sensitivity and specificity for the detection of ROP.3,8
The incidence of ROP in Mongolia has increased since 2000, when a study of visual impairment in Mongolia demonstrated that none of the patients were visually impaired due to ROP, which was attributed to the high rate of neonatal mortality at that time.11 Neonatal care in Mongolia has improved; the infant mortality rate in 2010 was approximately 12% less than in 2000.12 The rate of preterm birth in 2010 was 13.5%,13 indicating a significant population at risk of developing ROP.
The aim of this study was to describe a process for identifying BW and GA screening guidelines in Mongolia using telemedicine in the first reported cohort of ROP screening. Current screening guidelines for ROP at the National Center for Maternal and Child Health in Ulaanbaatar, Mongolia, include all infants with GA of younger than 34 weeks and BW of less than 2,000 g; these guidelines were established in accordance with the results from this study.
Patients and Methods
This study was conducted after obtaining approval from the Institutional Review Board at the University of Illinois at Chicago. Data were prospectively collected at a single center from December 1, 2015 to January 31, 2017. The Institutional Review Board approved the study protocol as not requiring parental consent due to the de-identified nature of the data, which were used for program monitoring purposes.
All neonates born at or referred to the National Center for Maternal and Child Health in Ulaanbaatar, Mongolia, during the study period who met inclusion criteria (GA ≤ 36 weeks and/or BW ≤ 2,000 g) entered a hospital screening protocol. Inclusion criteria were developed using a broad window based on international guidelines, including those developed in India by the National Neonatology Forum in 2010 and those produced by the Government of India's Ministry of Health and National Welfare in 2017.14,15 Dilated indirect ophthalmoscopy was performed using cyclopentolate 1% with an eyelid speculum and scleral depression. Indirect ophthalmoscopy and RetCam imaging (Natus Medical) were performed by a Mongolian ophthalmologist (TC).
Eyes were evaluated by a single, fellowship-trained Mongolian ophthalmologist (TC) with 3 years of ROP examination experience. Infants were screened for the presence or absence of ROP, zone of vascularization, stage, plus disease, and aggressive posterior ROP (AP-ROP). The diagnosis and classification of ROP for the current study was determined by examination using indirect ophthalmoscopy, and treatment plans were determined according to the International Classification for ROP and the Early Treatment for ROP (ET-ROP) Study.2,16 Type 1 ROP was defined as (1) zone I, any stage ROP with plus disease, (2) zone I, stage 3 ROP without plus disease, or (3) zone II, stage 2 or 3 ROP with plus disease. Type 2 ROP was defined as (1) zone I, stage 1 or 2 ROP without plus disease or (2) zone II, stage 3 ROP without plus disease. The infant was classified by the eye with the more severe disease.
Demographic factors, diagnosis, and clinical course were recorded in a de-identified manner using a ROP-specific REDCap database (Vanderbilt University). REDCap is a secure, web-based platform to collect image and demographic data. Demographic data collected included BW, GA, gender, and ethnicity/race. At each visit, the post-menstrual age, day of life, weight, oxygen status, and ROP examination details were input. RetCam images associated with each visit were linked to each survey. The data were entered by trained research staff to ensure accuracy. Training was performed on-site during the initial implementation and remotely if subsequent issues arose. Retinal imaging was performed using the RetCam and images were uploaded to the web-based platform, which could be accessed by both the Mongolian and United States ophthalmologists for two purposes: baseline imaging as a reference for follow-up examinations and verification of ROP grading by an international expert using a web-based platform.
Infants were categorized according to BW, GA, and zone, stage, and category of ROP as classified by the ET-ROP Study (no ROP, less than type 2 ROP, type 2 ROP, or type 1 ROP requiring treatment).2 BW and GA were categorized based on their quantile distribution as follows: BW (≤ 1,200 g, between 1,201 and 1,699 g, and ≥ 1,700 g) and GA (≤ 29 weeks, between 30 and 31 weeks, and ≥ 32 weeks). Visits with missing data were excluded from further analysis. Log-binomial regression analysis was performed to determine the association (relative risk with 95% confidence intervals) between BW and GA and ROP status, which represent dichotomous outcomes.
Data were collected from 196 infants, with seven examinations and three patients excluded from analysis due to missing data, leaving 329 examinations of 193 patients. The BW of infants in this study ranged from 750 to 2,000 g with a mean of 1,427 g, and the GA at birth ranged from 25 to 35 weeks with a mean of 30 weeks. There were 96 boys (49.7%) and 97 girls (50.3%). Among infants receiving treatment, the highest BW was 2,000 g (born at 31 weeks' GA) and the oldest was 34 weeks (with a BW of 1,300 g). Figure 1 demonstrates the distribution of BW and GA in the study cohort.
Distribution of birth weight (BW) and gestational age (GA) in the cohort of patients analyzed.
Current United States screening guidelines recommend that infants who are born at a GA of 30 weeks or younger and/or BW of 1,500 g or less be screened for ROP.3 A total of 18 infants (9%) had a BW and GA outside United States screening guidelines, 8 of whom had type 1 ROP. Of the 8 (4%) with type 1 ROP, 7 were treated with anti-vascular endothelial growth factor (VEGF) and 1 was treated with anti-VEGF and vitrectomy.
A total of 75 infants (38.9%) developed any stage of ROP and 44 infants (22.8%) developed type 2 or worse, whereas 4 infants (2.07%) progressed to stage 4 and none to stage 5. Of those with ROP, the highest severity reached was mild in 31 infants (41.3%), type 2 in 6 infants (8%), and type 1 in 38 infants (50%). The distribution of BW and GA of infants with ROP is shown in Table 1.
ROP Severity by BW and GA
Across all patients, a total of 38 patients were treated for ROP with either laser, anti-VEGF, or both anti-VEGF and vitrectomy. Thirty-two infants (84%) were treated with anti-VEGF only, 4 were initially treated with anti-VEGF and required subsequent vitrectomy (10%), and 2 were treated with laser only (6%). A summary of the types of ROP that required treatment is provided in Table 2. Of the 38 patients who were treated, 30 (79%) had zone I disease at some point in their treatment course. Of the 4 infants treated surgically, 2 (50%) had zone I disease at the time of surgical intervention. All treated patients had type 1 ROP, with 14 (36%) of the 38 treated patients having AP-ROP. Figure 2 demonstrates an example of a patient with AP-ROP who was treated with anti-VEGF.
Treated ROP by Category of Disease
An example of a patient who was treated with anti-vascular endothelial growth factor for aggressive posterior retinopathy of prematurity at post-menstrual age of 34 weeks. The patient's gestational age was 29 weeks and birth weight was 1,250 g. The image on the left is a temporal view and the image on the right is the inferior view. Both images are from the same patient.
The relative risk of developing both any stage of ROP and type 2 or worse ROP increases as the BW and GA decrease, which is demonstrated in Table 3. Infants with a BW of 1,200 g or less had 2.05 (95% CI: 0.99 to 4.25) times the risk of developing type 2 or worse ROP and 2.41 (95% CI: 1.35 to 4.29) times the risk of developing any type of ROP compared to those with a BW of 1,700 g or greater. The relative risk of developing any stage of ROP in infants with a GA of 29 weeks or younger was 2.91 (95% CI: 1.55 to 5.44) and a relative risk of 2.98 (95% CI: 1.22 to 7.28) of developing type 2 or worse ROP compared to those with a GA of 32 weeks or older, which represents a nearly three-fold difference.
Relative Risk of Developing Any Stage ROP or Type 2 or Worse ROP
The distributions of BW and GA among infants in this study developing ROP in Mongolia differ from those found in high-income countries, and are comparable to other low- and middle-income countries.10,11 The relative risk of developing ROP in our study cohort increased as BW and GA decreased. A web-based data management system was used effectively for collecting data in a pilot ROP program in a low- to middle-income country and this initial process was the foundation for the creation of an integrated tele-education, telemedicine, and data management platform. The results from this prospective study were used to develop evidence-based guidelines for ROP screening that are currently used in Mongolia.
The distribution of BW and GA among children affected by ROP in our study population from Mongolia are similar to those reported for other low- and middle-income countries,10,11,17–19 with similar ROP incidence at various BW and GA cut-offs20,21 and comparable mean BW and GA among affected children. However, the distributions of GA and BW among affected children differed significantly from those of high-income countries. Infants with heavier BW and older GA develop more severe ROP in low- and middle-income countries compared to high-income countries. Gilbert et al8 demonstrated that the mean BW of infants requiring treatment for threshold ROP from high-income countries ranged from 737 to 763 g, whereas low- and middle-income countries had a range of 903 to 1,527 g. They also demonstrated that the GAs of treated infants ranged from 25.3 to 25.6 weeks in high-income countries compared to 26.3 to 33.5 weeks in low- and middle-income countries.8 Notably, 18 infants with ROP, nearly one-third of those affected, had BW and GA that fell outside current United States ROP screening guidelines, and would have been missed if those criteria were applied in this setting. Our findings were similar to those found by Gilbert et al,8 who noted that 13% of infants in low- and middle-income countries exceeded United Kingdom screening requirements.
The greatest BW of an infant requiring treatment was 2,000 g and the greatest GA was 34 weeks. All infants requiring treatment would have been identified by screening infants with a GA of 34 weeks or less and/or a BW of 2,000 g or less. This prospective study provides information about the development of ROP in preterm infants referred to a tertiary institution in the capital city of Ulaanbaatar, Mongolia. The results of this study contributed to evidence-based screening guidelines that were subsequently implemented in Mongolia, as well as adding to the growing body of knowledge regarding trends in ROP in low- and middle-income countries affected by the epidemic of ROP.
It is important to note that although BW and GA are evidence-based predictors of ROP, there are likely other factors at play, and infants born outside those guidelines may develop ROP that requires treatment. Our study did not collect sufficient clinical data to evaluate these additional factors, but it should be noted that other screening guidelines include infants with an unstable clinical course who are outside the guidelines for BW and GA.3 Port et al22 described the profiles of outlier infants, described as infants with low BW who did not develop ROP and infants with greater BW who developed ROP, and found that the clinical factors typically associated with ROP development, including mechanical ventilation, multiple gestation, intraventricular hemorrhage, and race are not necessarily present in these “outlier infants.” We still do not have a definitive answer as to why some infants falling outside screening guidelines are affected by ROP. A case series by Padhi et al23 of larger and heavier infants who developed ROP in India proposed that fetal distress and an unstable neonatal course may play a role in inadequate peripheral retinal vascularization.
This study has several potential limitations. We only included data from a single tertiary referral center in an urban area. The patient population at this hospital likely represents a broader demographic given the density of the population in Ulaanbaatar relative to the entire country and referral patterns for higher acuity care in the capital city. Our study is similar to previous work characterizing infants with ROP in middle-income countries, which are also conducted in tertiary referral centers with similar limitations.17,18,24 The reliability of our analyses is dependent on the accuracy of estimated GA obtained from birth records. In 2017 at this tertiary referral center, 38% of preterm births occurred in rural areas and 0.6% of preterm births occurred at home. Limitations in the quality of data of BW and GA in low- to middle-income countries has been previously reported.25
It has been estimated that more than half of infants in low- to middle-income countries are never weighed at birth, especially those born at home, and facility-based data may have selection bias.26 The availability of ultrasound affects the assessment of gestational age, and estimation of GA based on last menstrual period has an estimated standard deviation error of 3 weeks.27
The severity of ROP was determined by a local ophthalmologist, which increases the applicability of our data in the local context. However, there may be variability in the diagnosis and staging of ROP that might have been mitigated with validation by expert graders. Multiple previous studies have demonstrated variability in grading of ROP by retina specialists and trainees from both the United States and internationally, particularly with regard to plus disease.24,28–32
Despite the potential limitations, this study and program established insight and guidance in determining screening criteria and program development for children at risk for ROP in Mongolia. In 2018, at the tertiary referral center, preterm births accounted for 13% of all births and 77% of the infants born prematurely weighed less than 2,500 g at birth. Therefore, it is possible that screening all infants with a BW of less than 2,000 g may place a high burden on providers responsible for screening for ROP, making telemedicine an attractive option for image-based screening. Current successful telemedicine programs have partnered academic institutions and tertiary referral hospitals with those that provide care in remote and rural regions.31–34 This is particularly applicable for ROP, where telemedicine can be used to consult internationally recognized experts in ROP diagnosis in challenging cases or in contexts where local providers are not adequately trained in ROP diagnosis.
Data management and standardization of data input remain a challenge, as noted in the limitations of this study. In the current study, data were captured on-site at the time of examination, but could be accessed and analyzed remotely using a web-based platform. This process has the potential to be replicated in other screening settings to obtain an international database specific to the various populations being screened. We used a secure, web-based data collection and retrieval system that could be extended to multiple countries. A similar process for developing evidence-based screening criteria could be used in other low- and middle-income countries, as well as aggregating data from additional, more remote screening centers in Mongolia to better characterize trends in ROP in the country. The web-based platform used in this study was the foundation for creating iTelegen, which is an integrated platform for telemedicine, data management, and web-based education modules.35 iTelegen has been used as a data management system and telemedicine platform for ROP screening in India33,34 and Nepal (unpublished data), and has also been used as a tele-education platform for modules in retinal vascular diseases such as diabetic retinopathy and ROP. This platform has the potential to not only address challenges in data management, but also provide continuing medical education for local providers.
The data from this study are consistent with previous studies that have demonstrated similar trends in BW and GA in low- and middle-income settings.10,11,17,18,20,21 We demonstrated that it is possible to successfully integrate a web-based platform for ROP screening as a tool for effective data management in the first reported cohort of infants who underwent a standardized ROP screening program in Mongolia.
- Palmer EA, Hardy RJ, Dobson V, et al. 15-year outcomes following threshold retinopathy of prematurity: final results from the multicenter trial of cryotherapy for retinopathy of prematurity. Arch Ophthalmol. 2005;123(3):311–318.
- Good WVEarly Treatment for Retinopathy of Prematurity Cooperative Group. Final results of the Early Treatment for Retinopathy of Prematurity (ETROP) randomized trial. Trans Am Ophthalmol Soc. 2004;102:233–248.
- Fierson WMAmerican Academy of Pediatrics Section on OphthalmologyAmerican Academy of OphthalmologyAmerican Association for Pediatric Ophthalmology and StrabismusAmerican Association of Certified Orthoptists. Screening examination of premature infants for retinopathy of prematurity. Pediatrics. 2013;131(1):189–195. doi:10.1542/peds.2012-2996 [CrossRef]
- Sommer A, Taylor HR, Ravilla TD, et al. Council of the American Ophthalmological Society. Challenges of ophthalmic care in the developing world. JAMA Ophthalmol. 2014;132(5):640–644. doi:10.1001/jamaophthalmol.2014.84 [CrossRef]
- Gilbert C. Retinopathy of prematurity: a global perspective of the epidemics, population of infants at risk and implications for control. Early Hum Dev. 2008;84(2):77–82. doi:10.1016/j.earlhumdev.2007.11.009 [CrossRef]
- Lad EM, Hernandez-Boussard T, Morton JM, Moshfeghi DM. Incidence of retinopathy of prematurity in the United States: 1997 through 2005. Am J Ophthalmol. 2009;148(3):451–458. doi:10.1016/j.ajo.2009.04.018 [CrossRef]
- Gilbert C, Rahi J, Eckstein M, O'Sullivan J, Foster A. Retinopathy of prematurity in middle-income countries. Lancet. 1997;350(9070):12–14. doi:10.1016/S0140-6736(97)01107-0 [CrossRef]
- Gilbert C, Fielder A, Gordillo L, et al. International NO-ROP Group. Characteristics of infants with severe retinopathy of prematurity in countries with low, moderate, and high levels of development: implications for screening programs. Pediatrics. 2005;115(5):e518–e525. doi:10.1542/peds.2004-1180 [CrossRef]
- Fierson WM, Capone A Jr, American Academy of Pediatrics Section on OphthalmologyAmerican Academy of OphthalmologyAmerican Association of Certified Orthoptists. Tele-medicine for evaluation of retinopathy of prematurity. Pediatrics. 2015;135(1):e238–e254. doi:10.1542/peds.2014-0978 [CrossRef]
- Shrestha JB, Bajimaya S, Sharma A, Shresthal J, Karmacharya P. Incidence of retinopathy of prematurity in a neonatal intensive care unit in Nepal. J Pediatr Ophthalmol Strabismus. 2010;47(5):297–300. doi:10.3928/01913913-20091118-08 [CrossRef]
- Bulgan T, Gilbert CE. Prevalence and causes of severe visual impairment and blindness in children in Mongolia. Ophthalmic Epidemiol. 2002;9(4):271–281. doi:10.1076/opep.9.4.271.1512 [CrossRef]
- Tsilaajav T, Ser-Od E, Baasai B, Byambaa G, Shagdarsuren O. Health systems in transition: Mongolia-health systems review. Geneva: World Health Organization; 2013.
- Blencowe H, Cousens S, Oestergaard MZ, et al. National, regional, and worldwide estimates of preterm birth rates in the year 2010 with time trends since 1990 for selected countries: a systematic analysis and implications. Lancet. 2012;379(9832):2162–2172. doi:10.1016/S0140-6736(12)60820-4 [CrossRef]
- Rashtriya Bal Swathya KaryakramMinistry of Health & Family Welfare, Government of India. Guidelines for Universal Eye Screening in Newborns Including Retinopathy of Prematurity. 2017. Accessed at: https://nhm.gov.in/images/pdf/programmes/RBSK/Resource_Documents/Revised_ROP_Guidelines-Web_Optimized.pdf
- “Retinopathy of Prematurity.”National Neonatology Forum of India Policy Guidelines. Accessed at: https://www.ontop-in.org/ontop-pen/Week-12-13/ROP%20NNF%20Guidelines%20.pdf
- International Committee for the Classification of Retinopathy of Prematurity. The international classification of retinopathy of prematurity revisited. Arch Ophthalmol. 2005;123(7):991.
- Beden U, Demir S, Aygün C, Küçüködük S, Erkan D. Screening for retinopathy of prematurity (ROP) in the middle Black Sea region of Turkey. Turk J Pediatr. 2012;54(3):223–229.
- Wu T, Zhang L, Tong Y, Qu Y, Xia B, Mu D. Retinopathy of prematurity among very low-birth-weight infants in China: incidence and perinatal risk factors. Invest Ophthalmol Vis Sci. 2018;59(2):757–763. doi:10.1167/iovs.17-23158 [CrossRef]
- Corpus K, Jimenez JM, Anzures R, Bascuna RI, Ventura R, Reandelar MRetinopathy of Prematurity Working Group. Proposed new retinopathy of prematurity screening criteria: evidence for including older and heavier Filipino premature babies. Philipp J Ophthalmol. 2013;38(2):72–79.
- Li ML, Hsu SM, Chang YS, et al. Retinopathy of prematurity in southern Taiwan: a 10-year tertiary medical center study. J Formos Med Assoc. 2013;112(8):445–453. doi:10.1016/j.jfma.2012.03.002 [CrossRef]
- Bas AY, Koc E, Dilmen UROP Neonatal Study Group. Incidence and severity of retinopathy of prematurity in Turkey. Br J Ophthalmol. 2015;99(10):1311–1314. doi:10.1136/bjophthalmol-2014-306286 [CrossRef]
- Port AD, Chan RV, Ostmo S, Choi D, Chiang MF. Risk factors for retinopathy of prematurity: insights from outlier infants. Graefes Arch Clin Exp Ophthalmol. 2014;252(10):1669–1677. doi:10.1007/s00417-014-2716-1 [CrossRef]
- Padhi TR, Rath S, Jalali S, et al. Larger and near-term baby retinopathy: a rare case series. Eye (Lond). 2015;29(2):286–289. doi:10.1038/eye.2014.253 [CrossRef]
- Vinekar A, Dogra MR, Sangtam T, Narang A, Gupta A. Retinopathy of prematurity in Asian Indian babies weighing greater than 1250 grams at birth: ten year data from a tertiary care center in a developing country. Indian J Ophthalmol. 2007;55(5):331–336. doi:10.4103/0301-4738.33817 [CrossRef]
- Lee AC, Katz J, Blencowe H, et al. CHERG SGA-Preterm Birth Working Group. National and regional estimates of term and preterm babies born small for gestational age in 138 low-income and middle-income countries in 2010. Lancet Glob Health. 2013;1(1):e26–e36. doi:10.1016/S2214-109X(13)70006-8 [CrossRef]
- Sreelatha OK, Ramesh SV. Teleophthalmology: improving patient outcomes?Clin Ophthalmol. 2016;10:285–295. doi:10.2147/OPTH.S80487 [CrossRef]
- Gunay M, Celik G, Tuten A, Karatekin G, Bardak H, Ovali F. Characteristics of severe retinopathy of prematurity in infants with birth weight above 1500 grams at a Referral Center in Turkey. PLoS One. 2016;11(8):e0161692. doi:10.1371/journal.pone.0161692 [CrossRef]
- Chan RVP, Williams SL, Yonekawa Y, Weissgold DJ, Lee TC, Chiang MF. Accuracy of retinopathy of prematurity diagnosis by retinal fellows. Retina. 2010;30(6):958–965. doi:10.1097/IAE.0b013e3181c9696a [CrossRef]
- Myung JS, Paul Chan RV, Espiritu MJ, et al. Accuracy of retinopathy of prematurity image-based diagnosis by pediatric ophthalmology fellows: implications for training. J AAPOS. 2011;15(6):573–578. doi:10.1016/j.jaapos.2011.06.011 [CrossRef]
- Campbell JP, Ryan MC, Lore E, et al. Imaging & Informatics in Retinopathy of Prematurity Research Consortium. Diagnostic discrepancies in retinopathy of prematurity classification. Ophthalmology. 2016;123(8):1795–1801. doi:10.1016/j.ophtha.2016.04.035 [CrossRef]
- Chiang MF, Jiang L, Gelman R, Du YE, Flynn JT. Interexpert agreement of plus disease diagnosis in retinopathy of prematurity. Arch Ophthalmol. 2007;125(7):875–880. doi:10.1001/archopht.125.7.875 [CrossRef]
- Wootton R. Telemedicine support for the developing world. J Telemed Telecare. 2008;14(3):109–114. doi:10.1258/jtt.2008.003001 [CrossRef]
- Valikodath N, Shah P, Chan RV, et al. Evaluation of birth weight and gestational age in infants with treatment requiring retinopathy of prematurity in ROPE-SOS trial. Invest Ophthalmol Vis Sci. 2018;59(9):2749.
- Valikodath N, Chan RV, Beca F, et al. Retinopathy of prematurity screening criteria based on the ROPE-SOS trial in India. Invest Ophthalmol Vis Sci. 2019;60(9):6522.
- Cole E, Beca F, Gil GL, et al. iTeleGEN: development of a web-based diabetic retinopathy education module. Invest Ophthalmol Vis Sci. 2019;60(9):5307.
ROP Severity by BW and GA
|Category||Mild||Type 2a||Type 1b||Total|
|BW (categorized) (g)|
| ≤ 1,200||13||6||15||34|
| 1,201 to 1,699||15||0||13||28|
| ≥ 1,700||3||0||10||13|
|GA (categorized) (wk)|
| ≤ 29||19||3||22||44|
| 30 to 31||8||3||12||23|
| ≥ 32||4||0||4||8|
Treated ROP by Category of Disease
|Category||Anti-VEGF Only (n = 32)||Anti-VEGF Followed by Vitrectomy (n = 4)||Laser Only (n = 2)|
|All Type 1b||32||4||2|
|Type 1 – AP-ROP||14 (44%)||2 (50%)||–|
Relative Risk of Developing Any Stage ROP or Type 2 or Worse ROP
|Category||Relative Risk (95% CI)|
|Any Stage ROP||Type 2a or Worse ROP|
|BW (categorized) (g)|
| 1,201 to 1,699 vs ≥ 1,700||2.0 (1.1 to 3.6), P = .02||1.4 (0.6 to 3.0), P = .41|
| ≤ 1,200 vs ≥ 1,700||2.4 (1.3 to 4.3), P = .003||2.1 (0.9 to 4.2), P = .05|
|GA (categorized) (wk)|
| 30 to 31 vs ≥ 32||2.1 (1.0 to 4.1), P = .035||2.4 (0.9 to 6.1), P = .02|
| ≤ 29 vs ≥ 32||2.9 (1.5 to 5.4), P = < .001||2.9 (1.2 to 7.3), P = .07|