The frequency of twin pregnancies has increased over the past two decades.1,2 In Denmark, the rate of twin pregnancies has more than doubled since the early 1980s.3 This trend is primarily due to an increasing use of assisted reproduction techniques and a higher maternal age. In 30–60% of multiple pregnancies, the neonates are born preterm, and, thus, multiple pregnancies constitute up to one fifth of all preterm deliveries. A neonate born preterm as a consequence of a multiple pregnancy has a higher mortality rate compared with a singleton neonate of the same gestational age.4,5 The cause of preterm delivery is multifactorial, involving a series of disorders and pathophysiologic mechanisms, but plausible explanations for preterm delivery after a multiple pregnancy is uterine overdistension and intrauterine infection.6,7 Other risk factors for preterm delivery include low and high maternal ages, nulliparity, previous preterm delivery, smoking, in vitro fertilization treatment (IVF), low socioeconomic status, and being single.1
A recently suggested risk factor for preterm delivery is loop electrosurgical excision procedure (LEEP), which is used worldwide as a treatment for cervical intraepithelial neoplasia (CIN), a cervical cancer precursor with a peak incidence in women in the reproductive age.8 In a previous study, we found a twofold increased risk of preterm delivery subsequent to LEEP among singletons.9
Thus, the aim of the present study was to investigate whether prior LEEP might be an additional risk factor for preterm delivery in a large population of twin deliveries. Furthermore, we aimed to investigate whether other cervical procedures (biopsy and ablation) are risk factors for spontaneous preterm twin delivery.
MATERIALS AND METHODS
From the population-based Medical Birth Registry and the National Patient Registry, we obtained information on all deliveries in Denmark over 9 years, 1997–2005. The unique personal identification number, used universally throughout the Danish society, was used in the linkage between registries. The study was found to be exempt by the Committees on Biomedical Research Ethics for the capital region of Denmark.
Details of this study have been described elsewhere.9 Briefly, we obtained data on 12,011 twin deliveries (24,022 neonates) with 21–45 completed weeks of gestation from the Medical Birth Registry and the National Patient Registry for the period January 1, 1997, through December 31, 2005. In the present study, we defined preterm delivery as a delivery between 21 and 37 completed gestational weeks. The data were linked to the Danish Registry of Pathology, from where we obtained information on all cervical procedures linked to the deliveries above (eg, date of the procedure and histology of the excised tissue). Information on IVF treatment was obtained from the IVF registry. Furthermore, obstetric history was obtained by linking all deliveries back to the Medical Birth Registry and the National Patient Registry a second time.
Information on all ablative procedures (local destruction of tissue by cauterization, laser or other erosive procedures) was obtained through the National Patient Registry. Before 1996, cold-knife conization and LEEP had the same code (LDC00) in the Nordic Classification of Surgical Procedures. Thus, to exclude cold knife conization, all deliveries with prior conization from before 1996 were excluded from the study population (n=111). Furthermore, we excluded deliveries subsequent to cold-knife conization from 1996 (n=6). We were not able to separate laser conization from LEEP, but the use of laser conization has since the introduction of LEEP in 1989 become very rare in Denmark. In addition, we excluded 431 deliveries where medical induction before 37 completed weeks had been performed, 1,593 deliveries where caesarean delivery had been performed before initiation of birth and before 37 completed weeks, and two deliveries subsequent to a collum amputation, to avoid inclusion of preterm deliveries of another origin. We compared the proportion of LEEP compared with no LEEP exposure in the above-mentioned excluded groups with the group of deliveries in the main analysis, but found virtually no differences between the two groups. Thus, the risk of introducing a competing risk bias was expected to be minimal. In all, 9,868 twin deliveries were eligible for further analysis, of which 3,228 were spontaneous preterm deliveries (32.7%).
Logistic regression analyses were used to evaluate the association between cervical procedures and preterm delivery, with simultaneous adjustment for potential confounders selected and applied on the basis of the literature, depending on the information available in the registries.1,6 Continuous variables included calendar time (year of delivery) and maternal age at delivery. Maternal age at delivery was modeled as a linear spline with knots placed at 20 and 40 years of age.10 Categorical variables included smoking during pregnancy (yes: stopped in first trimester; stopped after first trimester; less than 5, 6–10, 11–20, more than 20 cigarettes per day; amount unknown; no; missing data), IVF (yes, no), and marital status during pregnancy (no partner, no registered partner, married, unknown). Smoking and marital status during pregnancy were grouped according to the categories in the Medical Birth Registry. In addition, estimation was done with additional adjustment for obstetric history, specified as previous spontaneous abortion (yes, no), previous induced abortion (yes, no), and previous delivery (yes: only term, both term and preterm, only preterm; no; unknown). Furthermore, to minimize the effect of associations due to obstetric history, a secondary analysis regarding only primiparous women was performed. Effect modification by parity was investigated by allowing the estimated association between LEEP and preterm delivery to be different among primiparous and multiparous, respectively, and the estimates were compared using the generalized score statistic. Finally, we performed a secondary analysis in which pregnancies with the following characteristics were excluded: neonatal malformation, maternal infection, and maternal disease in pregnancy. Each of these characteristics covers a broad spectrum of conditions not otherwise specified, only some of which are potential risk factors.
Associations were expressed as odds ratios (ORs) with 95% confidence intervals (CIs). Generalized estimating equations using the exchangeable working correlation structure and robust variance estimates were applied to take into account correlation between the deliveries of one woman.11 Multinomial logistic regression was used to estimate the associations between cervical procedures and gestational age categorized into 21–27 weeks, 28–32 weeks, 33–36 weeks, and from 37 weeks (term), the last being the baseline category. Analyses were conducted with SAS 9.1 software (SAS Institute Inc, Cary, NC).
Of the 9,868 twin deliveries in the study, 166 were subsequent to LEEP (of which 11 were subsequent to both LEEP and ablation), 43 were subsequent to ablation, 766 were subsequent to biopsy with no additional cervical procedure, and the remaining 8,893 deliveries were not preceded by any cervical procedure. Only 4 (2.4%) of the deliveries subsequent to LEEP were preceded by more than one LEEP. The 9,868 deliveries were to 9,827 women, of whom none had more than two twin deliveries.
Table 1 shows the distribution of different risk factors by LEEP exposure and pregnancy outcome. In deliveries subsequent to LEEP, maternal age was slightly higher, and being single during pregnancy, smoking during pregnancy and being primiparous were more frequent. In preterm deliveries maternal age was slightly lower, marital status was more often cohabiting rather than married, and smoking during pregnancy was more frequent. Furthermore, primiparity and previous preterm delivery was more frequent in the preterm delivery group.
The association between LEEP and subsequent preterm delivery is presented in Table 2. Twin pregnancies subsequent to LEEP had a significantly increased risk of subsequent spontaneous preterm delivery (43.4%), with an adjusted OR of 1.58 (95% CI 1.16–2.14) compared with pregnancies with no prior LEEP (32.5%). The estimate did not change when comparing the LEEP group with the group of no prior procedures (data not shown). Further adjustment for obstetric history variables only attenuated the increased risk marginally (OR 1.55, 95% CI 1.14–2.12). Comparing deliveries subsequent to LEEP with deliveries subsequent to biopsy only, the risk was still significantly increased, with an adjusted OR of 1.54 (95% CI 1.09–2.19) (data not shown). Neither ablation nor biopsy before pregnancy were significantly associated with higher risks for spontaneous preterm delivery compared with pregnancies with no prior procedure (OR 1.05, 95% CI 0.56–1.98 and OR 1.02, 95% CI 0.87–1.19, respectively) (data not shown). Results from the analysis with preterm delivery grouped into three categories according to gestational age showed that the risk for spontaneous preterm delivery subsequent to LEEP was increased in all three categories of gestational age (although not significantly in moderately preterm delivery). The strength of the association between deliveries subsequent to LEEP and the two categories of very preterm and extremely preterm delivery were similar and higher than that for moderately preterm. Dividing the extremely preterm deliveries further into two categories of 21–23 weeks and 24–27 weeks revealed no changes in risk estimates, although this was based on small numbers (data not shown).
Estimating the risk for preterm delivery subsequent to LEEP in primiparous (n=4,945) and multiparous (n=4,923) compared with no LEEP deliveries yielded ORs of 1.33 (95% CI 0.89–1.99) and 1.86 (95% CI 1.16–3.00), respectively, revealing a nonsignificant difference between the two groups (P=.25). Excluding deliveries with certain pregnancy complications/characteristics (maternal infection, maternal disease, and neonatal malformation) did not alter the association in the remaining group of deliveries (n=7,252) (OR 1.66; 95% CI 1.14–2.42).
Our study shows that LEEP treatment of CIN is associated with a significant increase in risk for subsequent spontaneous preterm twin delivery, even after adjustment for a variety of confounding factors. When dividing preterm delivery into three categories according to gestational age, we find that the significant increase in risk for preterm delivery subsequent to LEEP is present in all three categories. In line with the singleton deliveries,9 the strongest association for preterm twin delivery is in the very preterm and extremely preterm groups.
The estimated associations between LEEP and preterm delivery were not significantly different for primiparous and multiparous deliveries; however, it was higher and only significantly elevated among multiparas. Comparing the risk for preterm delivery after LEEP in singleton pregnancies (OR 2.07, 95% CI 1.88–2.27) to twin pregnancies (OR 1.58, 95% CI 1.16–2.14), the estimates of the latter tended to be slightly lower; however, the difference was not statistically significant. Contributing to the tendency toward a lower risk for twin pregnancies may be that the twin group was relatively small, implying a lower precision of the estimate, and that twin pregnancy has an a priori much higher risk for preterm delivery than the singleton group.
Several previous studies comparing singleton deliveries with and without prior LEEP have shown a tendency toward an increased risk for preterm delivery8,12–14; however, no study has ever assessed the association between LEEP and risk of subsequent preterm delivery in twin pregnancies. Although the group of neonates with the greatest risk of morbidity and mortality comprises those born before 32 weeks of gestation, the group between 32 and 36 weeks still experiences an increased risk of health and developmental disorders compared with those born at term. Recently, some authors have used 34 weeks as cutoff week for preterm delivery, but as we present preterm delivery in subcategories, the traditional definition was used in this study. In the past decades, the increased frequency of twin deliveries has resulted in higher rates of preterm twin deliveries, but in contrast to singleton deliveries, the overall mortality rates have not decreased in twin deliveries.15,16 In addition to the higher risk of preterm delivery per se, twin pregnancies also differ from singleton pregnancies in that two infants are affected with the potential adverse consequences of a preterm delivery.
A recent study from Australia also showed that women with untreated CIN had a significantly higher risk for preterm delivery than women in the general population.17 In line with this, in a previous study, we introduced a group of singleton deliveries with prior biopsy and found that deliveries subsequent to biopsy only showed a slightly increased risk for preterm delivery compared with no prior procedure.9 In the present study, we did not find an increased risk for preterm delivery in the groups of twin deliveries subsequent to either biopsy or ablation treatment; however, the sample size in the latter group was small.
Our study has several strengths. First, the Danish health information system has an almost 100% coverage of the Danish population and a very high quality of the data. All deliveries in Denmark are registered online immediately after the delivery, ensuring minimal reporting and recall bias. Second, because the Danish Registry of Pathology receives data from every hospital and outpatient clinic in the country, we had virtually complete information on all surgical procedures performed before the deliveries in the study, thus ensuring a minimum of selection bias. Third, we adjusted for several potential confounding factors, including calendar time, maternal age, smoking during pregnancy, and marital status. Last, when performing an analysis including obstetric history variables, we found virtually no changes in the estimates.18
A potential limitation of the study is that the Medical Birth Registry does not contain information about socioeconomic status of the mother; therefore, this potentially confounding variable could not be included in the analyses. However, because both smoking during pregnancy and no partner in pregnancy to some extent reflect lower socioeconomic status in Denmark, we consider the residual confounding after adjusting for these variables to be limited.19 Another limitation is that, although phased out in the early 1990s, inclusion of a small number of laser conizations in the LEEP group is possible in our study. Finally, some of the categories in the stratified analysis were relatively small.
In conclusion, we find a significant increase in risk for spontaneous preterm delivery among twin pregnancies subsequent to LEEP, even after adjusting for several potential risk factors. The risk of preterm delivery subsequent to LEEP is especially strong for the very preterm and extremely preterm categories.
1. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet 2008;371:75–84.
2. Imaizumi Y. Trends of twinning rates in ten countries, 1972–1996. Acta Genet Med Gemellol (Roma) 1997;46:209–18.
3. Statistics Denmark. The Medical Birth Registry. Available at: http:www.sst.dk/English.aspx
. Retrieved June 26, 2009.
4. Slattery MM, Morrison JJ. Preterm delivery. Lancet 2002;360:1489–97.
5. Scher AI, Petterson B, Blair E, Ellenberg JH, Grether JK, Haan E, et al. The risk of mortality or cerebral palsy in twins: a collaborative population-based study. Pediatr Res 2002;52:671–81.
6. Romero R, Espinoza J, Kusanovic JP, Gotsch F, Hassan S, Erez O, et al. The preterm parturition syndrome [published erratum appears in BJOG 2008;115:674–5]. BJOG 2006;113 suppl:17–42.
7. Gardner MO, Goldenberg RL, Cliver SP, Tucker JM, Nelson KG, Copper RL. The origin and outcome of preterm twin pregnancies. Obstet Gynecol 1995;85:553–7.
8. Kyrgiou M, Koliopoulos G, Martin-Hirsch P, Arbyn M, Prendiville W, Paraskevaidis E. Obstetric outcomes after conservative treatment for intraepithelial or early invasive cervical lesions: systematic review and meta-analysis. Lancet 2006;367:489–98.
9. Noehr B, Jensen A, Frederiksen K, Tabor A, Kjaer SK. Loop electrosurgical excision of the cervix and subsequent risk for spontaneous preterm delivery: a population-based study of singleton deliveries during a 9-year period. Am J Obstet Gynecol 2009 Apr 3. [Epub ahead of print].
10. Harrel FE Jr. Regression modeling strategies. New York (NY): Springer-Verlag; 2001. p. 16–26.
11. Zeger SL. Liang KY, Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121–30.
12. Sadler L, Saftlas A, Wang W, Exeter M, Whittaker J, McCowan L. Treatment for cervical intraepithelial neoplasia and risk of preterm delivery. JAMA 2004;291:2100–6.
13. Samson SL, Bentley JR, Fahey TJ, McKay DJ, Gill GH. The effect of loop electrosurgical excision procedure on future pregnancy outcome. Obstet Gynecol 2005;105:325–32.
14. Jakobsson M, Gissler M, Sainio S, Paavonen J, Tapper AM. Preterm delivery after surgical treatment for cervical intraepithelial neoplasia [published erratum in Obstet Gynecol 2008;112:945]. Obstet Gynecol 2007;109:309–13.
15. Joseph KS, Allen AC, Dodds L, Vincer MJ, Armson BA. Causes and consequences of recent increases in preterm birth among twins. Obstet Gynecol 2001;98:57–64.
16. Kogan MD, Alexander GR, Kotelchuck M, Macdorman MF, Buekens P, Martin JA, et al. Trends in twin birth outcomes and prenatal care utilization in the United States, 1981–1997. JAMA 2000;284:335–41.
17. Bruinsma F, Lumley J, Tan J, Quinn M. Precancerous changes in the cervix and risk of subsequent preterm birth. Br J Obstet Gynaecol 2007;114:70–80.
18. Petersen ML, Sinisi SE, van der Laan MJ. Estimation of direct causal effects. Epidemiology 2006;17:276–84.
19. Wisborg K, Henriksen TB, Hedegaard M, Secher NJ. Smoking habits among Danish pregnant women from 1989 to 1996 in relation to sociodemographic and lifestyle factors. Acta Obstet Gynecol Scand 1998;77:836–40.