A Cross-Sectional Study of Obstetrical and Gynecological Parameters in Twins – A Potential Increase in Premature Rupture of Membranes and Pre-eclampsia in Twin Women

Background This cross-sectional study of 192 twins across a wide geographical area compared obstetrical and gynecological parameters between twins and the general population. Methods The study is a cross-sectional study. Gynecological variables included age of menarche, abnormal pap smears, birth control usage, endometriosis, broids, infertility, gynecological cancer, and osteoporosis. Obstetrical variables studied included pregnancy, ectopic pregnancies, hypertension, pre-eclampsia, gestational diabetes, and premature rupture of membranes (PROM). Chi Square and Fisher’s Exact tests were utilized to compare these binary variables between identical and fraternal twins. Finally, odds ratios were used to compare the rates of PROM, pre-eclampsia, hypertension, ectopic pregnancy, and gestational diabetes among twins compared to the general population using reputable literature sources.


Conclusions
Our ndings suggest that further large scale research concerning PROM and pre-eclampsia in twin women may be warranted.

Background
Twin pregnancies have long been a topic of study and investigation. One has little di culty nding an abundance of information on the diagnosis, management, complications and delivery recommendations for twin gestations. There have been tremendous advances in monochorionic twin pregnancy management and therapy up to and including the role of fetoscopic laser coagulation of placental anastomoses for Twin-Twin Transfusion Syndrome management. What happens when these very twins become of childbearing age however? Are they more likely to develop pre-eclampsia themselves or experience premature rupture of membranes? Would they be more likely to develop gestational diabetes than those born of singleton pregnancies? If twins would be more likely to develop pre-eclampsia or gestational diabetes should they routinely be offered aspirin for pre-eclampsia prophylaxis or early glucose tolerance testing when they become pregnant? The questions remain unanswered. Interestingly, data to answer such questions is nonexistent. Given the increased incidence of premature rupture of membranes, pre-eclampsia, and gestational diabetes in twin gestations it seems feasible that these twin women could be at increased risk of these conditions when they themselves become pregnant.
The gynecological questions surrounding twins are equally unknown. Are identical or fraternal twins more likely to both have the same age of menarche, abnormal pap smears, use birth control, have endometriosis, have broids, experience infertility, develop gynecological cancer, or be diagnosed with osteoporosis? The goals of this study were to investigate some of these questions. More speci cally, social, obstetrical, and gynecological relationships between female identical and fraternal twins were investigated. Additionally, obstetrical frequencies among twins were compared to those in the general population.

Methods
This study was fully approved by the Institutional Review Board of Albany Medical Center/Albany Medical College before its commencement. The study was a cross sectional study involving the distribution of surveys to female twins. The surveys were developed speci cally for this study. In order to be eligible to participate in the study twins needed to be female, be an identical or fraternal twin, and be at least 18 years of age. Surveys were given in registration packets to all twins who attended the Twins Days Festival in Twinsburg, Ohio, in August of 2014. Additionally, surveys were electronically mailed to all twins on the Twins Days Festival mailing list which included twins from the festival's inception in 1976. Of note, the Twins Days Festival is the largest annual gathering of twins and multiples in the world. Twins were instructed to complete the surveys independently of their twin counterpart and not to collaborate with their twin in any way. Participation in this survey was completely optional and twins could readily opt out of participating. Once a survey was completed it was mailed back to the Department of Obstetrics and Gynecology at Albany Medical Center in Albany, New York. Each survey couplet had a unique tracking number on it so the surveys could be matched back together upon receival.
The contents of the individual surveys contained demographic, social, gynecological, and obstetrical questions. Questions were constructed in a binary manner to elicit a yes/no response. Results were tabulated and coded as binary variables either yes (1) or no (0). SAS University Edition was then utilized to perform all statistical analyses. Speci cally, Chi Square and Fisher's Exact Tests were used to compare all binary variables. To compare obstetrical ratios in our twin sample to the general population a literature review was performed and odds ratios were calculated. The incidence of premature rupture of membranes in the general population was estimated to be 10% of all pregnancies [1], the incidence of pre-eclampsia was estimated to be 4% of all pregnancies [2], the incidence of hypertension was estimated to be 10% of all pregnancies [3], the incidence of ectopic pregnancies was estimated to be 2% of all pregnancies [4], and the incidence of gestational diabetes was estimated to be 6% of all pregnancies [5].
Using these population estimates, it was determined that an adequate sample size was obtained for all the above variables in this survey to have a con dence level of 95% that the real value is within ± 5% of the surveyed value with each individual population proportion. (Needed sample size of 139 for PROM, 60 for Pre-eclampsia, 139 for hypertension, 31 for ectopic, 87 for gestational diabetes: all of which were obtained). Finally, in order to calculate odds ratios to the general population the number of pregnancies and births in the United States both had to be estimated. Recent CDC data was utilized to estimate the annual number of births in the United States as 3,791,712 [6] and the annual number of pregnancies as 6,155,000 [7].

Results
In total 192 surveys were received representing 96 sets of twins. Using the attendance at the festival alone, in which there is estimated to be 2,000 female twins annually, the response rate was estimated to be 9.6%. When factoring in the large number of surveys that were sent electronically, the response rate was likely much smaller than this. In total, surveys were received from 23 states, Canada, and Great Britain. 41.7% were from Eastern States, 40.6% were from the Midwest, 9.4% were from Southern States, 6.3% were from Western States, and 1% were from Canada and Great Britain respectively ( Figure I). 172 of the twins identi ed as identical twins and 20 identi ed as fraternal twins. The average age of respondents was 43.3 years with a standard deviation of 17.0 years ( Figure II). Of the respondents, 87.5% identi ed as White/Hispanic, 8.3% identi ed as Black, 2.1% identi ed as Asian, and 2.1% identi ed as other. This racial breakdown is similar to the United States general population in which it is estimated that 78% are White/Hispanic, 12% are Black, 6% are Asian, and 4% identify as other [8]. Social factor results are identi ed in Table I. Gynecological factors are identi ed in Table II and obstetrical factors in   Table III. There was a total of 258 reported pregnancies. A statistically signi cant greater number of identical twins reported having the same age of menarche than did fraternal twins (69.51% verse 33.33%, p-value: 0.0021). Other gynecological parameters did not have a statistically signi cant difference between identical and fraternal twins (Table II). There were no statistically signi cant differences detected among the obstetrical parameters tested between identical and fraternal twins (Table III) Table IV. Since pregnant women reported their total number of pregnancies and if they ever had a given obstetrical complication but did not report in how many of their pregnancies a complication occurred, two separate odds ratios were calculated. The rst odds ratio was calculated assuming that only one of their pregnancies was affected (Minimum Odds Ratio) and the second odds ratio was calculated assuming all of their pregnancies were affected (Maximum Odds Ratio). The true odds ratio undoubtedly lies between these two calculations. The minimum and maximum odds ratios are shown in Table V. Using these calculations, there was an increase between 37% and 442% in the odds of PROM and 17-216% in the odds of pre-eclampsia in our twin sample verse the general population

Discussion
The increased incidence of PROM and pre-eclampsia in our twin sample verse the general population was of great interest. A literature review identi ed the incidence of PROM in the general population to be approximately 10% [1], whereas 24.3% of our pregnant women reported undergoing PROM in at least one of their pregnancies. Likewise, the incidence of pre-eclampsia in the general population is estimated to be about 4% [2], whereas 8.4% of our pregnant women reported having pre-eclampsia in at least one of their pregnancies. Since the number of pregnancies effected for each twin who had the condition was not known however, minimum and maximum odds ratios were calculated. Using this technique, the minimum odds ratio for PROM was 1.37 (95% CI = 0.95,1.96) and the maximum odds ratio was 5.42 (95% CI = 4.21,6.98). The actual odds ratio undoubtedly lies between these two numbers and is most likely signi cant.
Using the same technique the minimum odds ratio for pre-eclampsia was 1.17 (95% CI = 0.66,2.09) and the maximum odds ratio was 3.16 (95% CI = 2.16,4.62) in our twin sample verse the general population.
Since both PROM and pre-eclampsia are more likely in subsequent pregnancies when one pregnancy has been affected, it is likely that at least some of the women in our sample were affected in more than one pregnancy. Interestingly, twin gestations themselves are associated with an increased risk of both PROM and pre-eclampsia. Whether there is a causative role behind the increased risk of PROM and preeclampsia during the antecedent twin gestation and subsequent pregnancies of these twin women remains unclear but is a potential area of further research. For instance, if twin women are more likely to develop pre-eclampsia than singleton women, aspirin prophylaxis for twin women may be routinely offered. Although there is wide variation on the usage of aspirin for pre-eclampsia prophylaxis, The American College of Obstetricians and Gynecologists (ACOG) does not currently list being a twin as an indication for prescribing aspirin for pre-eclampsia prophylaxis. Further, since the incidence of PROM and pre-eclampsia is known to vary across race, the racial breakdown in our sample needed to be similar to the general United States population for meaningful comparisons to be made. Overall, the socioeconomic factors that make up this twin cohort are similar to the general United States population. The racial distribution of our twin sample was similar to the general population with 87.5% verse 78% identifying as White/Hispanic, 8.3% verse 12% identifying as Black, 2.1% verse 6% identifying as Asian, and 2.1% verse 4% identifying as other in our sample verse the general population respectively [8].
Another important consideration was the impact of smoking and PROM in our sample. Since smoking is a known cause of PROM it was very important to consider the potential role of tobacco use in our twin sample verse the general population. It has been estimated that in 2005, 13.8% of women reported smoking during pregnancy [9]. In our twin sample 9.1% of pregnant women reported smoking at some point during their life. Whether the tobacco use in the twin women occurred during pregnancy or not was not speci ed so the actual percentage using tobacco during pregnancy was likely lower than 9.1%. Certainly, while smoking is often under-reported the smaller percentage in our twin sample verse the general population makes it unlikely that tobacco use played a signi cant role in our PROM ndings.

Page 6/10
The scant knowledge of amnionicity and chorionicity among twins was also striking. Although this is one of the most critical factors in the management and prognosis of twin gestations, less than 1 in 10 (9.9%) of twins sampled had this knowledge about their pregnancy. Interestingly, no fraternal twins (0/18) reported having this knowledge. However, every single twin who responded to this survey identi ed as either an identical or fraternal twin. Since 80 percent of all twins are dichorionic [2], and could thus be either identical or fraternal, the fact that 100% of the sample reportedly knew they were either identical or fraternal was curious. At any rate, there appears to be a general lack of knowledge about the embryology of twin pregnancies among twins themselves.
There were both strengths and limitations to this study. Strengths included a diverse group of female twins in terms of both age and demographics and completely voluntary participation in the study. The cross sectional study design allowed multiple obstetrical and gynecological parameters to be analyzed at the same time. Finally, the medical literature was useful in providing comparisons to our study sample. A signi cant limitation of the study was that there was no way to de nitively corroborate that the twins took the surveys independently despite instructions to not collaborate. Unfortunately, this resulted in the possibility of twin collaboration and respondent bias. There was also a signi cant possibility of recall bias especially among older twins recounting obstetrical issues. The response rate was also lower than desired. Additionally, there were many non-statistically signi cant results and more identical twins than fraternal twins responded to the survey.

Conclusions
This study suggests a potential increase in PROM and pre-eclampsia in twin women, although further large scale research is necessary to further investigate if such a relationship exists. Further areas of research could include large scale analysis of PROM and pre-eclampsia in twin women and potential alterations in counseling and management of these pregnancies.  Questionnaire.docx