LAUSR

Infections with human papillomaviruses (HPVs) are a necessary cause of cervical cancer, which affects over 500000 women worldwide every year. Beyond that, HPV infections may be associated with negative reproductive outcomes like preterm birth (1). The causal relationship of cervical HPV infections and their sequelae with obstetric outcomes is not well understood. Potential underlying causes include genetic predisposition, direct effects of HPV on obstetric outcomes, and factors related to treatment of cervical precancers. Effects of treatment have been most widely studied, and data suggest that treatment can lead to cervical insufficiency and obstetric complications with a documented dose–response relationship between depth of excision and risk for preterm birth. However, evaluating direct effects of HPV on obstetric outcomes is difficult because HPV infection cannot be easily separated from progression to precancer and treatment effects. Furthermore, other sexually transmitted infections, as well as socioeconomic and lifestyle factors associated with HPV, can confound observed associations and are often not sufficiently documented. Changes in screening, management practices, and treatment approaches over time can affect historical comparisons, further limiting the ability to infer causal associations. In their article, He and colleagues (2) describe elevated risks for preterm birth, chorioamnionitis, infant sepsis, and neonatal death among obstetric deliveries in Swedish women after a diagnosis of cervical intraepithelial neoplasia grade 3 (CIN3) compared with both their female siblings and women in the general population of Sweden. They stratified the analysis by calendar period to account for changing approaches to CIN treatment and also compared spontaneous and iatrogenic preterm birth. They noted a decrease in the severity of outcomes over time, attributed to practice change away from coldknife conization, which removes more cervical tissue, to the loop electrosurgical excision procedure, which generally removes less cervical tissue. Although receipt, timing, and type of treatment in relationship to the index pregnancy could not be verified in this study, it was assumed that treatment rates were high and treatment practices were consistent within each time period. The findings of He and colleagues are consistent with previous findings that described increased rates of adverse obstetric outcomes among women treated for CIN3, especially as excision depth increases (3–5). However, the study could not address the challenging question of whether increased pregnancy complications are the result of HPV infections or other environmental or biological factors that predispose to CIN3, CIN3 itself as a manifestation of HPV infection, CIN3 treatments, or a combination of these factors. To address the possibility of a genetic predisposition underlying observed associations between HPV-related disease and obstetric outcomes, He and colleagues compared obstetric outcomes in women with previous CIN3 and their siblings without previous CIN3. This analysis showed similar results to the comparison with the overall population, suggesting that genetic predisposition does not explain the observed associations. Research comparing women with CIN3 with the general population generally indicates higher rates of pregnancy complications, independent of whether CIN3 was treated. For example, Shanbhag and colleagues (6) found that women diagnosed with CIN3 had a 52% higher risk for preterm delivery than women without a CIN diagnosis, and this risk was independent of treatment type. Castanon and colleagues (7) examined whether preterm birth occurred more frequently before or after a diagnosis of CIN. They found that women with CIN were at higher risk for preterm birth than the general population but noted no effect of treatment after adjusting for timing of diagnosis relative to pregnancy (7). Conversely, Werner and colleagues (8) examined rates of preterm delivery among women seeking care at a large safety-net hospital with an overall preterm birth rate of 7%. In their population, neither CIN nor the loop electrosurgical excision procedure conferred increased risk for preterm delivery; they found no difference in the likelihood of preterm birth between women undergoing loop electrosurgical excision for CIN treatment and the general hospital population regardless of whether the treatment occurred before or after the index pregnancy (8). Because of variable and sometimes conflicting study conclusions regarding the relationship of CIN3 and its treatment to subsequent obstetric complications, conclusions about their relative roles remain uncertain. A recent Cochrane review concluded that published results should be interpreted with caution because of low or very low data quality (3). In summary, most studies indicate that women with CIN3 are at risk for preterm birth and adverse obstetric outcomes, but the underlying reasons and optimal management strategies for preventing pregnancy complications continue to be uncertain. Although He and colleagues' study design addresses possible genetic factors, it could not further disentangle the effects of treatment from those of CIN3 or HPV itself, which limits the ability to inform clinical care. The likely multifactorial cause of increased obstetric risk in those with a CIN3 diagnosis, whether treated or untreated, may never be fully proven or understood. Because CIN3 is a direct cancer precursor, treatment should always be provided to avoid the development of invasive cancer unless the patient is currently pregnant (9). Although this study suggests that pregnancies after treatment of CIN3 be managed as “high risk” to decrease adverse outcomes, the optimal clinical care remains unclear. We currently lack specific obstetric interventions or care algorithms to mitigate the risk for pregnancy complications associated with CIN3 diagnoses. Excision