Auger, Nathalie MD, MSc; Fraser, William D. MD, MSc; Paradis, Gilles MD, MSc; Healy-Profitós, Jessica MPH; Hsieh, Ada; Rhéaume, Marc-André MD
Retinal disease is a major contributor to blindness and reduced visual acuity. Globally, an estimated 90 eyes are blinded per hour by retinal detachment alone.1 Risk factors for retinal disorders are not fully understood.2 Characteristics such as aging, diabetes, and myopia explain only part of the disease burden associated with the retina.3–7 Pregnancy-related risk factors, however, have received little attention. This is especially true for preeclampsia, which is known to affect the retina during pregnancy.8 Women with preeclampsia are at risk for central serous chorioretinopathy,9 serous retinal detachment,10 and retinopathy during pregnancy,9 but the period extending past delivery has received less consideration.
Preeclampsia is an understudied risk factor for long-term retinal disease. Preeclampsia is relatively common, occurring in 3–5% of pregnant women, and is an important cause of maternal morbidity.11–13 Preeclampsia is characterized by hypertension with proteinuria or evidence of other organ involvement after 20 weeks of gestation.11 Most of the retinal changes in women with preeclampsia resolve after delivery or within a few months postpartum,8,10 but the extent to which preeclampsia can affect the retina beyond this point is less understood. Recent evidence suggests that preeclampsia is associated with an increased risk of cardiovascular and renal disease in the long term.12,14 The growing evidence of lifelong implications of preeclampsia leads to the question of whether an association also exists with later retinal disease. Our objective therefore was to evaluate the relationship between preeclampsia and risk of retinal detachment or retinopathy in the decades after pregnancy.
MATERIALS AND METHODS
We carried out a longitudinal cohort study of all women who delivered neonates in any hospital in the province of Quebec, Canada, between 1989 and 2013. We tracked the women longitudinally over time for any hospitalization after delivery with follow-up ending March 31, 2014. In Quebec, 99% of women deliver in a hospital.15 We therefore had near complete coverage of all pregnancies in the province. We restricted the study to women with pregnancies that reached at least 20 weeks of gestation, because preeclampsia occurs rarely before this time. To carry out the study, we extracted all data from hospital discharge summaries in the Maintenance and Use of Data for the Study of Hospital Clientele registry. Discharge summaries in Quebec are completed by trained hospital personnel, validated using rigorous algorithms, and contain up to 26 diagnostic and 15 procedural codes for each admission, including death during hospitalization.16
We identified women with preeclampsia during any pregnancy using the International Classification of Diseases (ICD), 9th and 10th Revisions. We categorized preeclampsia according to severity with ICD codes for mild (642.3, 642.4, O13) or severe or superimposed (642.5–642.7, O11, O14, O15). The ICD differentiates mild and severe preeclampsia based on degree of proteinuria. During the study, clinically significant proteinuria was defined as 300 mg or greater collected during a 24-hour period (mild preeclampsia) and heavy proteinuria as 3,000 g or greater per day (severe preeclampsia).17 Superimposed preeclampsia consists of new-onset proteinuria with pre-existing hypertension. We included gestational hypertension in the same category as mild preeclampsia for its potential to be on the spectrum of preeclampsia.18 The 10th Revision of the ICD does not separate mild preeclampsia and gestational hypertension.
We further distinguished preeclampsia based on time of onset, including early (less than 34 weeks of gestation) and late (34 weeks of gestation or greater).11,13 Early- and late-onset preeclampsia are believed to have separate etiologies with potential for differing relationships with retinal disorders later in life.13
We examined a range of outcomes, including retinal detachment (full or partial separation of retinal tissue from the back of the eye), retinopathy (persistent inflammation and vascular remodeling of the retina), and other retinal disorders. Examples of other disorders included retinal vascular occlusion, retinal hemorrhage, and macular degeneration. Retinal detachments were categorized as serous, traction, rhegmatogenous, breaks, and unspecified and retinopathy as diabetic and nondiabetic. We identified all retinal disorders using diagnostic codes in the ICD or procedure codes in the Canadian Classification of Diagnostic, Therapeutic, and Surgical Procedures and the Canadian Classification of Health Interventions (Appendix 1, available online at http://links.lww.com/AOG/A896).2 Because the data were hospital-based, we primarily identified retinal outcomes severe enough to require treatment or be documented at hospitalization. In Quebec, procedures for retinal detachment are mostly carried out in a hospital, facilitating capture of women who ever required treatment. Mild serous detachments, retinal breaks, and retinopathy may be treated in outpatient facilities.3
We had several covariates as potential confounders or mediators of the relationship between preeclampsia and retinal disease, including baseline age (younger than 20, 20–24, 25–29, 30–34, 35–39, 40 years or older), parity (one, two, three or more deliveries), socioeconomic deprivation (poorest fifth of the population, no, unknown),19 and period (1989–1996, 1997–2004, 2005–2013) at cohort entry. We identified women with metabolic disease (diabetes, hypertension, obesity, dyslipidemia) at any point during the study using ICD codes (Appendix 1, http://links.lww.com/AOG/A896). Diabetes and hypertension in particular are risk factors for several retinal disorders5,6,20 and may be mediators. Women with preeclampsia are at high risk of developing hypertension and diabetes later in life.12
We calculated the cumulative incidence of retinal disorders for women with and without preeclampsia accounting for death as a competing outcome.21 We used Cox proportional hazards regression to estimate hazard ratios and 95% confidence intervals for the association between preeclampsia and later risk of retinal disorders. We began by analyzing preeclampsia as a single exposure without considering severity. We subsequently ran models according to severity of preeclampsia and gestational age of onset. We used the number of days since the index delivery as the time scale and censored women who died or had no event by the last day of follow-up, March 31, 2014.21 We verified the proportional hazards assumption using interaction terms with time.
We tested models that were unadjusted; partially adjusted for age, parity, socioeconomic deprivation, and time period; and fully adjusted for all covariates including metabolic disease (diabetes, hypertension, obesity, dyslipidemia). We used the partially adjusted model to assess whether characteristics of women such as age and deprivation were important confounders and the fully adjusted model to identify any remaining effect of preeclampsia on risk of retinal disorders not mediated through metabolic disease.
In sensitivity analyses, we assessed whether associations were influenced by women whose preeclampsia status was potentially misclassified, particularly at the start of the study as a result of incomplete pregnancy history. To do so, we reran regression models after excluding women with deliveries before 1996. We reran models excluding women with pre-existing hypertension or diabetes and those who developed retinal disorders during pregnancy. Finally, we tested models adjusted for metabolic diseases individually.
We used SAS 9.3 for data analysis. We used two-sided hypothesis tests and set statistical significance at P=.05. We focused on the magnitude and direction of the associations, because relying solely on statistical significance to guide interpretation is discouraged.22 The University of Montreal Hospital Centre institutional review board waived ethical review because the data were deidentified and the study conformed to ethical requirements.
In this study, 1,108,541 women delivered at least one neonate between 1989 and 2013. There were 64,350 women with preeclampsia (5.8%): 44,967 with mild and 19,383 with severe. There were 5,034 women with early-onset (0.5%) and 59,316 with late-onset preeclampsia (5.3%). During the 16,121,590 person-years of follow-up after delivery, 1,283 (0.12%) women were hospitalized for retinal detachment and 627 (0.06%) for retinopathy. Women with preeclampsia had proportionately more retinal disorders and metabolic disease compared with no preeclampsia (Table 1). In addition, women with preeclampsia were on average younger at the index pregnancy.
Preeclamptic women had higher cumulative incidence of all retinal outcomes compared with nonpreeclamptic women (Fig. 1). Compared with no preeclampsia, women with preeclampsia had a higher incidence of retinal detachment (52.9 compared with 23.9/10,000 women) and retinopathy (60.5 compared with 8.0/10,000). The incidence of any retinal detachment or retinopathy in women with preeclampsia increased early and steadily after the index delivery, but for serous or traction detachments and breaks, the incidence diverged from no preeclampsia after 15 years of follow-up. There was little evidence, however, of a difference in the incidence of rhegmatogenous detachments between women with and without preeclampsia.
Cumulative incidence...Image Tools
In unadjusted models, preeclamptic women had 5.3 times the risk of traction detachments and 3.7 times the risk of retinal breaks (Table 2). Risks were elevated but to a lesser degree for other types of retinal detachments, including rhegmatogenous and serous. Adjustment for age, parity, period, and socioeconomic status had little influence on the general magnitude of associations, indicating that these covariates were not major confounders. Additional adjustment for metabolic disorders significantly attenuated the associations, particularly for traction detachments and diabetic retinopathy. Even with adjustment for all covariates, risks associated with preeclampsia continued to be elevated for traction detachments, retinal breaks, and diabetic retinopathy.
Table 2-a. Associati...Image Tools
When preeclampsia was evaluated by severity, incidence of retinal disorders was higher for women with severe than mild preeclampsia (Table 3). Compared with no preeclampsia, the risk of any retinal detachment was 2.5 times higher for women with severe preeclampsia but only 34% higher for mild preeclampsia. The gradient with severity of preeclampsia persisted regardless of adjustment for metabolic disease. Women with severe preeclampsia had four times the risk of traction detachments and retinal breaks, whereas women with mild preeclampsia had only twice the risk compared with no preeclampsia. Similarly, severe preeclampsia was more strongly associated with diabetic and nondiabetic retinopathy than mild preeclampsia, although the difference was not as large.
When evaluated by gestational age of onset, risk of diabetic and nondiabetic retinopathy was higher among women with early- than late-onset preeclampsia (Table 4). Compared with no preeclampsia, women with early-onset preeclampsia had 8.4 times the risk of diabetic retinopathy and 4.6 times the risk of nondiabetic retinopathy. Associations for late-onset preeclampsia were weaker with 3.6 times the risk of diabetic retinopathy and 1.9 times the risk of nondiabetic retinopathy compared with no preeclampsia. In contrast, the risk of any retinal detachment was similar between early- and late-onset preeclampsia. The small number of women with early-onset preeclampsia prevented analysis of different types of retinal detachment.
Table 4-a. Gestation...Image Tools
In sensitivity analysis excluding women who gave birth before 1996, we found no change in the magnitude of associations. Excluding women with pre-existing diabetes or hypertension lowered the risk of diabetic retinopathy by 50% (hazard ratio 2.15, 95% confidence interval 1.40–3.32, preeclampsia compared with no preeclampsia), but had no effect on other outcomes. Excluding women with retinal disorders during pregnancy decreased the magnitude of associations slightly, but without changing the interpretation. Finally, adjusting for each metabolic disease individually did not change the results.
Table 4-b. Gestation...Image Tools
In this longitudinal study, women with preeclampsia, particularly severe and early-onset preeclampsia, had elevated risks of retinal disease in the decades after pregnancy. The incidence of most retinal disorders after pregnancy diverged almost immediately between women with and without preeclampsia. Metabolic diseases such as diabetes and hypertension attenuated the associations, but not enough to fully explain the relationship. These findings align with another study that found similar relationships in a smaller sample of Israeli women, but for ophthalmologic disorders overall.23 Our findings suggest that preeclampsia is associated with long-term risk of retinal disorders, that is, beyond the first year postpartum.9 Women with preeclampsia may benefit from increased vigilance for retinal disorders after pregnancy.
Serous retinal disorders are some of the more common problems in preeclampsia.9,10 Serous retinal detachment is thought to be secondary to hypertensive vasoconstriction and retinal pigment epithelium damage.10 In our cohort, severe preeclampsia was associated with greater risk of serous retinal detachment, but much of this was explained by metabolic disorders, which mediated the associations.
Risks were more elevated for traction detachment and retinal breaks. Traction detachment is thought to involve retinal ischemia with overexpression of angiogenic growth factors stimulating fibrosis.6 The pathways are reminiscent of preeclampsia, which also involves imbalanced growth factor production and pathologic angiogenesis.24 Preeclamptic women in our cohort had a high risk of traction detachments, even after adjustment for diabetes, a primary risk factor for traction detachment.6 Growing evidence that angiogenic imbalance persists after pregnancy suggests that preeclampsia may reflect a pathophysiologic profile with chronic undetected low-grade angiogenic dysfunction,25,26 manifesting as greater risk of traction detachments long term. The same pathways may be involved in the pathogenesis of cardiovascular diseases in preeclamptic women.12,14
Although preeclampsia was not associated with rhegmatogenous retinal detachments, there was an association with breaks, which typically precede rhegmatogenous detachments.3,4 The number of women affected was however low, and mild breaks managed in outpatient clinics were not included. Nevertheless, the greater risk of breaks and stronger association of rhegmatogenous detachment with severe than mild preeclampsia raises the possibility of a link. Preeclampsia could hypothetically weaken the retina during pregnancy, increasing the propensity for small breaks that go undetected or that develop postpartum.
Women with preeclampsia had a significantly greater risk of retinopathy, even after adjusting for metabolic disease.5 The pathogenesis of retinopathy is not fully understood, and data on how diabetes and hypertension interact with preeclampsia are limited. In a study of 158 Finnish women with type 1 diabetes, preeclampsia was associated with twice the risk of diabetic retinopathy 16 years after pregnancy.27 Preeclampsia may be an independent risk factor for diabetic and nondiabetic retinopathy with part of the effect mediated through diabetes or hypertension. Early-onset preeclampsia was even more strongly associated with retinopathy. These findings align with evidence that early- and late-onset preeclampsia reflects different pathologies,13 a distinction that may be more important for retinopathy than retinal detachment.
Our study has several limitations. We used administrative hospital data, and nondifferential coding errors may be present. International Classification of Diseases diagnostic criteria for preeclampsia have changed over time, potentially introducing errors of inclusion and exclusion that are difficult to identify. International Classification of Diseases codes limited our ability to identify milder gestational hypertension or end-organ dysfunction, including any organ affected. We accounted for diabetes and hypertension as time-fixed covariates, because we did not have the exact time of onset. We captured retinal disorders treated in the hospital, not outpatient clinics. We had limited information on the circumstances surrounding retinal disorders, including blindness and visual acuity. Similarly, we did not have information on smoking, ethnicity, or drug use and cannot completely rule out residual confounding. A proportion of women who were not hospitalized may have moved from the province, but this is unlikely to vary between preeclampsia and no preeclampsia. This was a Canadian population-based study and generalizability to other populations remains to be established.
In this population-based study of more than 1 million women with follow-up extending over two decades, preeclampsia was associated with a moderate risk of hospitalization for retinal disease later in life. The greatest risk was for traction detachments, retinal breaks, and retinopathy. Metabolic diseases such as diabetes and hypertension explained only part of the relationship, suggesting that preeclampsia could be an independent risk factor for retinal disorders. The American Academy of Ophthalmology recommends yearly screening of diabetic and hypertensive patients, but preeclampsia is not included.28 Further research is needed to determine whether a history of preeclampsia should also be an indication for screening.
1. Shah S. Blindness and visual impairment due to retinal diseases. Community Eye Health 2009;22:8–9.
2. Mitry D, Charteris DG, Fleck BW, Campbell H, Singh J. The epidemiology of rhegmatogenous retinal detachment: geographical variation and clinical associations. Br J Ophthalmol 2010;94:678–84.
3. D'Amico DJ. Clinical practice. Primary retinal retachment. N Engl J Med 2008;359:2346–54.
4. Sodhi A, Leung LS, Do DV, Gower EW, Schein OD, Handa JT. Recent trends in the management of rhegmatogenous retinal detachment. Surv Ophthalmol 2008;53:50–67.
5. Venkatramani J, Mitchell P. Ocular and systemic causes of retinopathy in patients without diabetes mellitus. BMJ 2004;328:625–9.
6. Eliott D, Hemeida T. Diabetic traction retinal detachment. Int Ophthalmol Clin 2009;49:153–65.
7. Chen SN, Lian IB, Wei YJ. Epidemiology and clinical characteristics of rhegmatogenous retinal detachment in Taiwan. Br J Ophthalmol 2016;100:1216–20.
8. Roos NM, Wiegman MJ, Jansonius NM, Zeeman GG. Visual disturbances in (pre)eclampsia. Obstet Gynecol Surv 2012;67:242–50.
9. Errera MH, Kohly RP, da Cruz L. Pregnancy-associated retinal diseases and their management. Surv Ophthalmol 2013;58:127–42.
10. Gundlach E, Junker B, Gross N, Hansen LL, Pielen A. Bilateral serous retinal detachment. Br J Ophthalmol 2013;97:939–40, 949.
11. Mol B, Roberts C, Thangaratinam S, Magee L, de Groot C, Hofmeyr G. Pre-eclampsia. Lancet 2016;387:999–1011.
12. Chen CW, Jaffe IZ, Karumanchi SA. Pre-eclampsia and cardiovascular disease. Cardiovasc Res 2014;101:579–86.
13. von Dadelszen P, Magee L. Pre-eclampsia: an update. Curr Hypertens Rep 2014;16:454.
14. Brown M, Best K, Pearce M, Waugh J, Robson SC, Bell R. Cardiovascular disease risk in women with pre-eclampsia: systematic review and meta-analysis. Eur J Epidemiol 2013;28:1–19.
15. Auger N, Le TU, Park AL, Luo ZC. Association between maternal comorbidity and preterm birth by severity and clinical subtype: retrospective cohort study. BMC Pregnancy Childbirth 2011;11:67.
16. Ministère de la Santé et des Services sociaux. Cadre normatif du système Med-Écho. (Maintenance et exploitation des données pour l'étude de la clientèle hospitalière). Québec, Québec: Gouvernement du Québec 2016:1–259.
17. Magee LA, Helewa M, Moutquin JM, von Dadelszen P; Hypertension Guideline Committee; Strategic Training Initiative in Research in the Reproductive Health Sciences (STIRRHS) Scholars. Diagnosis, evaluation, and management of the hypertensive disorders of pregnancy. J Obstet Gynaecol Can 2008;30(suppl):S1–48.
18. Melamed N, Ray JG, Hladunewich M, Cox B, Kingdom JC. Gestational hypertension and preeclampsia: are they the same disease? J Obstet Gynaecol Can 2014;36:642–7.
19. Auger N, Fraser WD, Healy-Profitós J, Arbour L. Association between preeclampsia and congenital heart defects. JAMA 2015;314:1588–98.
20. Farioli A, Hemmingsson T, Kriebel D. Vascular risk factors and rhegmatogenous retinal detachment: a follow-up of a national cohort of Swedish men. Br J Ophthalmol 2015 Oct 15 [Epub ahead of print].
21. Lau B, Cole SR, Gange SJ. Competing risk regression models for epidemiologic data. Am J Epidemiol 2009;170:244–56.
22. Stang A, Poole C, Kuss O. The ongoing tyranny of statistical significance testing in biomedical research. Eur J Epidemiol 2010;25:225–30.
23. Beharier O, Davidson E, Sergienko R, Szaingurten-Solodkin I, Kessous R, Charach R, et al.. Preeclampsia and future risk for maternal ophthalmic complications. Am J Perinatol 2016;33:703–7.
24. Steinberg G, Khankin EV, Karumanchi SA. Angiogenic factors and preeclampsia. Thromb Res 2009;123(suppl 2):S93–9.
25. Staff AC, Redman CW, Williams D, Leeson P, Moe K, Thilaganathan B, et al.. Pregnancy and long-term maternal cardiovascular health progress through harmonization of research cohorts and biobanks. Hypertension 2016;67:251–60.
26. Tuzcu ZB, Asicioglu E, Sunbul M, Ozben B, Arikan H, Koc M. Circulating endothelial cell number and markers of endothelial dysfunction in previously preeclamptic women. Am J Obstet Gynecol 2015;213:533.e1–7.
27. Gordin D, Kaaja R, Forsblom C, Hiilesmaa V, Teramo K, Groop PH. Pre-eclampsia and pregnancy-induced hypertension are associated with severe diabetic retinopathy in type 1 diabetes later in life. Acta Diabetol 2012;50:781–7.
28. American Academy of Ophthalmology. Frequency of ocular examinations—2015. 2015. Available at: http://http://www.aao.org
/clinical-statement/frequency-of-ocular-examinations. Retrieved June 21, 2016.
© 2017 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.