|Home | About | Journals | Submit | Contact Us | Français|
Presented in part: 15th International Workshop on Observational Databases, Prague, Czech Republic, 24 March 2011.
Background.Use of antiretroviral drugs (ARVs) during pregnancy has been associated with higher risk of preterm birth.
Methods.The Pediatric HIV/AIDS Cohort Study network's Surveillance Monitoring for ART Toxicities study is a US-based cohort of human immunodeficiency virus (HIV)–exposed uninfected children. We evaluated maternal ARV use during pregnancy and the risk of any type of preterm birth (ie, birth before 37 completed weeks of gestation), the risk of spontaneous preterm birth (ie, preterm birth that occurred after preterm labor or membrane rupture, without other complications), and the risk of small for gestational age (SGA; ie, a birth weight of <10th percentile for gestational age). Multivariable logistic regression models were used to evaluate the association of ARVs and timing of exposure, while adjusting for maternal characteristics.
Results.Among 1869 singleton births, 18.6% were preterm, 10.2% were spontaneous preterm, and 7.3% were SGA. A total of 89% used 3-drug combination ARV regimens during pregnancy. In adjusted models, the odds of preterm birth and spontaneous preterm birth were significantly greater among mothers who used protease inhibitors during the first trimester (adjusted odds ratios, 1.55 and 1.59, respectively) but not among mothers who used nonnucleoside reverse-transcriptase inhibitor or triple-nucleoside regimens during the first trimester. Combination ARV exposure starting later in pregnancy was not associated with increased risk. No associations were observed between SGA and exposure to combination ARV regimens.
Conclusions.Protease inhibitor use early in pregnancy may be associated with increased risk for prematurity.
Combination antiretroviral (ARV) regimens started early in pregnancy are associated with a reduction in maternal-to-child transmission of human immunodeficiency virus (HIV) to <1% . Combination ARV regimens (≥3 ARVs) are thus recommended for pregnant women worldwide who have CD4+ T-lymphocyte counts of <350 cells/mm3 . However, the potential for combination regimens to increase the risk of preterm birth is unclear. Increases in preterm birth could contribute significantly to infant morbidity and mortality in low-resource settings.
Early European studies suggested an increase in preterm birth with combination ARV use during pregnancy, especially when started before pregnancy [3, 4]. However, US data from the same period did not show an increased risk overall with combination regimens . A meta-analysis of 14 studies found no increase in preterm birth with use of ARVs, compared with no use of ARVs, but detected an increase in preterm delivery for those receiving combination regimens with protease inhibitors (PIs), compared with those using non-PI regimens, and for those receiving combination regimens before pregnancy or in the first trimester, compared with those who started combination regimens later in pregnancy . Subsequent studies have continued to yield conflicting results regarding whether combination regimens, and PIs specifically, are associated with increased risk of preterm birth [7–15]. Potentially least subject to confounding are data from a randomized trial in Botswana, which showed a nearly 2-fold higher preterm birth rate among women randomly assigned to receive a combination regimen including lopinavir/ritonavir, compared with those randomly assigned to receive a triple-nucleoside regimen .
In view of expanded access to combination ARV regimens for pregnant women worldwide and the need to understand the conditions most likely to be associated with preterm delivery, we evaluated maternal ARV exposure and other risk factors for preterm birth in the Pediatric HIV/AIDS Cohort Study (PHACS). The large size of the study, which includes detailed information on the specific type and timing of ARV regimens and other potential risk factors in pregnancy, allowed us to control for many potential confounders and to assess combinations of factors that may influence the risk of preterm delivery.
We analyzed data from HIV-infected pregnant women and their infants who were enrolled into the Surveillance Monitoring for ART Toxicities (SMARTT) study of the PHACS network. The SMARTT study includes 2 cohorts. The static cohort enrolled women and their children <12 years of age during 2007–2009 into ongoing surveillance; all subjects had previously participated in the International Maternal Pediatric and Adolescent AIDS Clinical Trials (protocol P1025 [for pregnant women] or 219C [for infants]), in the Women and Infants Transmission Study, or in other established local cohorts that met the minimum data collection requirements, including detailed information on antiretroviral use during pregnancy. The dynamic cohort, which began in 2007, enrolled pregnant women from 22 weeks of gestation through 1 week after birth and their infants through 1 week after birth into ongoing prospective surveillance. All protocols were approved by the appropriate local institutional review board, and all women provided written informed consent for study participation. Detailed information was collected via medical chart abstraction, by study nurses with perinatal research expertise, on maternal ARV use during pregnancy, during genital infections, and during other medical conditions; information regarding risk behaviors such as smoking and drug and alcohol use was collected as described below. Data on pregnancy and infant outcomes were abstracted from hospital medical records and obtained at a study-related newborn examination for those in the dynamic cohort. Both static cohort and dynamic cohort participants are followed prospectively with annual study visits.
Our analysis was limited to singleton gestations with maternal enrollment on or before 31 October 2010. Gestational age was based on best obstetrical estimate, determined using clinical findings in 43% of cases, ultrasound dating in 20%, and both ultrasound and clinical findings in 29%. Preterm birth was defined as delivery before 37 completed weeks of gestation and very preterm birth as delivery before 32 completed weeks of gestation. Preterm births were coded as “spontaneous” if they occurred after preterm rupture of membranes or with a diagnosis of preterm labor without other conditions, such as preeclampsia, and were vaginal or cesarean after labor or rupture of membranes . Preterm births were coded as “indicated” if they occurred in the presence of a maternal indication for delivery, such as preeclampsia, or in the presence of a fetal indication for delivery, such as fetal distress or growth restriction, and involved either cesarean section, without labor and ruptured membranes, or induction . A computer-based algorithm was developed for coding to ensure consistency across cases. Small for gestational age (SGA) was defined as a birth weight of <10th percentile for gestational age .
Start and stop dates for all ARVs taken during pregnancy were obtained and used to code trimester(s) of exposure to each agent . Combination ARV regimens were defined as those containing ≥3 ARVs, including drugs from ≥2 classes, or as those containing ≥3 nucleoside reverse-transcriptase inhibitors (NRTIs). Data on income, education level, smoking, alcohol, and drug use during pregnancy were collected by interview. Illicit drug use included use of marijuana, cocaine, opiates, methamphetamines, or phencyclidine. Self-reported data on drug use were validated in a subset of women on the basis of results of meconium testing . Information on history of maternal genital infections during pregnancy, including Neisseria gonorrhoeae infection, Chlamydia trachomatis infection, Trichomonas vaginalis infection, and Treponema pallidum infection, was abstracted from prenatal records. The latest CD4+ T-lymphocyte counts and percentages and HIV RNA levels available prior to labor and delivery were abstracted from medical records, as were the earliest corresponding measurements during pregnancy, if available.
Demographic and maternal characteristics were summarized overall and compared between those with preterm birth versus term birth, using Fisher exact tests, Pearson χ2 tests, or Wilcoxon rank sum tests, as appropriate. Trends in preterm birth rates over time were evaluated using Mantel-Haenszel tests for trend. Logistic regression models were used to evaluate the association of any type of preterm delivery, spontaneous preterm delivery, and SGA with in utero ARV exposures, both unadjusted and adjusted for potential confounders, including maternal HIV disease measures, genital infections, demographic characteristics, clinical site, and risk behaviors (alcohol, tobacco, and illicit drug use). Indicated births were excluded from the analysis of spontaneous preterm births because they may reflect preexisting maternal conditions not associated with ARV exposure, but they were included in the analysis of any type of preterm births because of the possibility that some indications for preterm delivery could themselves be associated with ARV exposure. Covariates with a P value of < .15 in univariate analyses were considered for inclusion in multivariate models for each outcome. The multivariate model was reduced to a final core model for each outcome that included covariates with a P value of < .10. ARV exposures were evaluated by drug class, individual drug, and trimester. Models were developed for evaluating risk of preterm birth for infants exposed to combination ARV including a PI, combination ARV with an NNRTI, and combination ARV with ≥3 NRTIs (but not a PI or NNRTI), compared with ARV monotherapy or dual therapy. We also evaluated the risk of preterm birth on the basis of first trimester exposure to each of these regimens, using women who were not receiving combination ARV regimens in the first trimester (ie, women receiving ARV monotherapy or dual therapy during the first trimester or combination ARV after the first trimester) as the reference category. Analyses of both overall and first trimester regimens excluded 59 infants born to mothers who received no ARV during pregnancy, since these women were generally not in prenatal care and often lacked other maternal health information. Because this was a multicenter study with potential variability in populations served by different research sites, a sensitivity analysis was conducted to adjust for research site. Sensitivity analyses were also conducted using random effects models to control for repeated deliveries in the same woman. Analyses were conducted using SAS, version 9.2 (SAS Institute, Cary, NC), and 2-sided P values of < .05 were considered statistically significant.
Of 2218 maternal/infant pairs enrolled into SMARTT as of October 2010, 215 were excluded because incomplete ARV or obstetric data during pregnancy, usually because of recent delivery, because of continuing gestation, or because permission had not yet been obtained for release of P1025 study data. An additional 95 infants were excluded because they were not singletons, and 39 were excluded because data on gestational age were missing, leaving 1869 live singleton infants born to 1506 women. Mean birth weight and gestational age, when available, for infants born to women with incomplete ARV data were very similar to those of infants included in our analysis. Characteristics of women, overall and by presence or absence of preterm birth, are shown in Table Table1.1. Of the 1869 infants in the study (including 70 infants who were known to be born before term but lacked data on exact gestational age), 346 (18.6%) were born before term. Of these preterm births, 191 (55%) were classified as spontaneous, 153 (44%) were classified as indicated, and 2 (0.6%) could not be classified. Very preterm delivery occurred in 37 of 1799 deliveries (2.1%); given this small number, risk factors for this outcome were not evaluated in detail. SGA occurred among 135 of 1861 infants (7.3%) who had information on both birth weight and gestational age. Combination ARV use was reported by 89% of women during pregnancy, with 40% reporting first-trimester use of combination ARV. Among infants who were not exposed to combination ARV during pregnancy, 70% were exposed to ARV monotherapy or dual therapy, and the remaining 30% were exposed to no ARVs. Among mothers who were not receiving first-trimester combination ARV regimens, 63% received combination ARVs during the second trimester, and an additional 13% received combination ARVs in the third trimester.
The distribution of infants by birth cohort did not differ significantly between preterm and full-term births (P = .32). However, the percentage of infants born before term showed a potentially important increasing trend over birth cohorts (from 15.7% for births before 2002 to 19.7% for births during 2008–2010; P = .08). This trend was no longer apparent after adjustment for race, maternal CD4+ T-cell count, and income level (P = .16).
Factors that were associated with an increased risk of any type of preterm birth (ie, spontaneous or indicated) in univariate analyses included black (ie, African or African American) race, income below $20 000/year, CD4+ T-lymphocyte count of <200 cells/mm3 late in pregnancy, and HIV RNA level of >1000 copies/mL late in pregnancy (Table (Table2).2). These factors all remained significant in adjusted models, except for HIV RNA level of >1000 copies/mL late in pregnancy. When we evaluated risk factors for spontaneous preterm birth in univariate models, these same factors, as well as illicit drug use and genital infections, were associated with this outcome. On multivariate analysis, only black race, income below $20 000/year, and CD4+ T-lymphocyte count of <200 cells/mm3 late in pregnancy were significantly associated with spontaneous preterm birth.
After initial models were developed, the type and timing of ARVs were evaluated in models that adjusted for race, income, and maternal CD4+ T-cell count at delivery. In unadjusted models, PI-based combination regimens were associated with a marginally increased risk of any type of preterm birth (OR, 1.60; 95% CI, .96–2.67; P = .07), compared with ARV monotherapy or dual therapy, but use of combination regimens with an NNRTI (OR, 1.33; 95% CI, .71–2.58; P = .36) or with ≥3 NRTIs (OR, 1.13; 95% CI, .60–2,14) was not. After adjustment for maternal health status and socioeconomic factors, there was no association of any combination ARV regimen with preterm birth, compared with ARV monotherapy or dual therapy. Combination ARV regimens with or without PIs were not associated with spontaneous preterm birth (Table (Table22).
However, when evaluating the timing of combination ARV, first-trimester combination regimens including a PI were associated with an increased risk of any type of preterm birth and of spontaneous preterm birth, with adjusted ORs of 1.55 to 1.59, respectively (Table (Table3).3). Estimated associations between any type of preterm birth and either first-trimester exposure to combination ARV or exposure to specific drug classes remained essentially unchanged in sensitivity analyses that controlled for multiple pregnancies per woman by using a random effect model and in models that adjusted for site effects. Preterm birth occurred among 155 of 748 women (21%) with first-trimester exposure to combination ARVs and among 155 of 924 women (17%) with initial exposure to combination ARV during the second or third trimester (P = .043, by the Fisher exact test). Very preterm birth occurred among 3% of those with first-trimester exposure and among 1% of those with later exposure to combination regimens (P = .005, by the Fisher exact test). In addition, a significantly elevated risk of very preterm birth was observed with combination regimens including PIs in the first trimester, in both unadjusted analysis (OR, 3.33; 95% CI, 1.53–7.27; P = .002) and adjusted analysis (adjusted OR, 4.17; 95% CI, 1.70–10.26; P = .003), compared with no combination regimen in the first trimester. Despite the differences in proportions with preterm birth, the median birth weights between the 2 groups were not significantly different, at 2990 g in the group with first trimester exposure and 3010 g in the group with no first trimester exposure (P = .49, by the Wilcoxon rank sum test; data not shown).
To evaluate whether the immunologic response to ARVs was related to any type of preterm birth, we assessed change in CD4+ T-lymphocyte percentages during pregnancy among women delivering at full term, compared with those delivering before term. The mean maternal CD4+ T-cell percentage increase was 3.1% for women with full-term births, compared with 2.2% for women with preterm births (P = .005). The mean percentage change per week was 0.20% for full-term births and 0.07% for preterm births (P = .07). Note that women who reported use of combination ARVs during the first trimester tended to have slightly higher CD4+ T-lymphocyte percentages and lower HIV-1 plasma RNA concentrations early in pregnancy than those with use first reported later during pregnancy (median CD4+ T-cell percentage, 27% vs 26%; percentage of women with a viral load of > 1000 copies/mL, 41% vs 68%).
Among women with ARV use during pregnancy, we next evaluated the association between individual drugs and preterm delivery, using adjusted models. Any preterm delivery was associated with use of saquinavir (adjusted OR, 2.32; 95% CI, 1.26–4.27; 4% exposed), ritonavir (adjusted OR, 1.35; 95% CI, 1.03–1.77; 48% exposed), and lopinavir/ritonavir (adjusted OR, 1.32; 95% CI, 1.00–1.74; 32% exposed), and similar associations were observed for saquinavir and ritonavir with spontaneous preterm birth. In addition, the risk for spontaneous preterm delivery was significantly decreased with zidovudine use (adjusted OR, 0.62; 95% CI, .41–.92; 83% exposed) and was marginally decreased with use of nelfinavir (adjusted OR, 0.68; 95% CI, .45–1.05; 28% exposed) and lamivudine (adjusted OR, 0.70; 95%CI, .47–1.04; 82% exposed). All analyses were adjusted for black race, income below $20 000/year, and maternal CD4+ T-lymphocyte count of <200 cells/mm3 prior to labor or delivery.
Finally, we observed no association of SGA with combination ARV exposure at any time during pregnancy or early in pregnancy, nor differences by first trimester exposure to individual drug classes (Table (Table4).4). Models for SGA were adjusted for the same covariates as those related to preterm birth and spontaneous preterm birth and had similar adjusted ORs, except that tobacco use during pregnancy was found to be associated with SGA (adjusted OR, 1.54; 95% CI, .96–2.47; P = .07) and was also included in the adjusted models. In addition, while black race was significantly associated with SGA, its effect appeared protective (adjusted OR, 0.65; 95% CI, .43– 1.00; P = .048), in contrast to the association of black race with increased risk of preterm birth.
We found that PI-based combination regimen use in the first trimester was associated with a risk of any type of preterm birth and spontaneous preterm birth. The magnitude of the risk was similar for any type of preterm birth and spontaneous preterm birth, suggesting that preterm birth in HIV-infected women is a heterogeneous condition, as it is in HIV-uninfected women. Several other studies have reported an association between continuing combination regimens from before pregnancy and increased risk of preterm birth, although not all studies adjusted for maternal disease stage [4, 6, 11]. Taken together, these results imply that there may be residual confounding, even after adjustment for maternal CD4+ T-lymphocyte count, suggesting that HIV disease progression or effects of combination ARV on the immune system among women with indications for initiation of therapy before pregnancy may contribute to increasing risk of preterm birth.
While we found a significant increase in preterm birth with the use of PI-containing combination ARV regimens in early pregnancy, the median birth weights were similar between the 2 exposure groups, suggesting that most of the increased risk is for later preterm birth. This finding is consistent with results of the randomized trial in Botswana, which found a doubling of preterm birth with PI therapy but similar median gestational age at delivery and no differences in neonatal hospitalizations or death by 6 months of age . We did not have detailed information on neonatal hospital stays to evaluate this factor for differences, but neonatal complications by type of ARV regimen are important outcomes for evaluation in future studies. In addition, we found a significant increase in preterm births before 32 weeks of gestation with early use of PI-containing combination ARV regimens, which requires further surveillance. Births at lower gestational ages are associated with higher mortality and long-term morbidity rates .
We found no association with overall use (ie, at any time during pregnancy) of PI-containing, NNRTI-containing, or triple NRTI combination ARV regimens with risk of any type of preterm delivery or of spontaneous preterm delivery. Instead, factors associated with an increased risk of preterm birth in the general population, including black race and low income, were associated with increased risk of any type of preterm birth and spontaneous preterm birth . In addition, CD4+ T-lymphocyte counts of <200 cells/mm3 were associated with both outcomes, suggesting that advanced HIV disease increases the risk of preterm birth. Low CD4+ T-lymphocyte count has been associated with preterm birth in the eras before and after combination ARV regimens emerged and in the United States and Europe [4, 7, 13, 15, 22, 23].
A recent study from Spain found an increased risk of spontaneous preterm birth among HIV-infected women as compared to matched HIV-uninfected women, and risk was increased with severe immune suppression . Interestingly, the Spanish study also found an increased risk of iatrogenic, or indicated, preterm birth with combination therapy started in the second half of pregnancy . Immune reconstitution with improved control of HIV infection may produce immunologic changes that increase the risk of adverse pregnancy outcomes . We found a modest association with better improvement in CD4+ T-lymphocyte percentage and decreased risk of preterm birth, arguing against immune reconstitution as a contributor. However, a limitation of our analysis is that we did not always have CD4+ T-lymphocyte measurements available prior to initiation of ARV regimens during pregnancy. Women who reported use of combination ARV during the first trimester tended to have slightly higher CD4+ T-lymphocyte percentages and lower HIV-1 plasma RNA concentrations than those with use first reported later during pregnancy, suggesting that the increased risk of preterm birth was not simply due to their more compromised health status. Our analyses should be considered exploratory and, thus, directed at suggesting future directions for further research.
We speculate that other immunologic and inflammatory mechanisms may play a role in the initiation of preterm labor. Interleukin 10 (IL-10), an antiinflammatory cytokine produced by monocytes and macrophages, is thought to be crucial for the maintenance of pregnancy . Studies have shown that the IL-10 level is higher among untreated HIV-infected subjects and that levels decrease ARV therapy and rebound after discontinuing therapy . Thus, combination ARV regimen use before and early during pregnancy may reduce IL-10 levels, interfering with maintenance of pregnancy. A small study suggested associations between decreasing IL-10 production by peripheral blood mononuclear cells in vitro in combination ARV–treated women and preterm delivery . In the same study, elevated interleukin 2 levels were marginally associated with preterm delivery. Further study is required to evaluate whether changes in circulating proinflammatory and antiinflammatory cytokine levels among pregnant women receiving combination regimens are associated with the risk of preterm birth.
We did not find any association with combination ARV use or with individual drug exposures during pregnancy and the risk of SGA. This finding is consistent with previous studies among infants exposed in utero to combination ARVs [28, 29].
While our data are generally reassuring, the association between first trimester PI combination drug exposure and increased preterm birth risk, including very preterm birth, raises concerns that warrant further study. In addition, although based on small numbers, the association between ritonavir use and spontaneous preterm birth is consistent with recent studies, including a randomized trial suggesting an increased risk of preterm birth with ritonavir-containing regimens. This emphasizes the need for additional data, especially given expanding access to combination drug regimens internationally [16, 30]. The ongoing Promoting Maternal and Infant Survival Everywhere study, which randomly assigns women with CD4+ T-lymphocyte counts of >350 cells/mm3 to a lopinavir/ritonavir-based combination regimen or to zidovudine during pregnancy with additional intrapartum drugs will provide important information regarding the optimal regimen for prevention of perinatal transmission, taking into account transmission rates, pregnancy outcomes, and long-term maternal health. In addition, appropriate specimens are being collected for further testing of immune and inflammatory mediators among HIV-infected women starting ARVs during pregnancy and their association, if any, with pregnancy outcome.
In conclusion, we have observed that although combination ARV use later in pregnancy is not associated with an increased risk of preterm delivery, use in the first trimester of PI-containing combination ARV regimens may contribute to an increased risk. The mechanism of first trimester effect is unclear but could be related to changes in immune and inflammatory mediators. Further studies are required to elucidate the specific drug effects and the interaction of the many factors that determine pregnancy outcome.
Acknowledgments.We thank the children and families who participated in PHACS and the individuals and institutions involved in the conduct of PHACS.
Data management services were provided by Frontier Science and Technology Research Foundation (principal investigator, Suzanne Siminski), and regulatory services and logistical support were provided by Westat (principal investigator, Julie Davidson). The following institutions, clinical site investigators, and staff participated in conducting PHACS SMARTT in 2010, in alphabetical order: Baylor College of Medicine: William Shearer, Mary Paul, Norma Cooper, Lynette Harris; Bronx Lebanon Hospital Center: Murli Purswani, Emma Stuard, Anna Cintron; Children's Diagnostic and Treatment Center: Ana Puga, Dia Cooley, Doyle Patton, Deyana Leon; Children's Hospital of Philadelphia: Richard Rutstein, Carol Vincent, Nancy Silverman; Children's Memorial Hospital: Ram Yogev, Margaret Ann Sanders, Kathleen Malee, Scott Hunter; Jacobi Medical Center: Andrew Wiznia, Marlene Burey, Molly Nozyce; New York University School of Medicine: William Borkowsky, Sandra Deygoo, Helen Rozelman; St. Jude Children's Research Hospital: Katherine Knapp, Kim Allison, Megan Wilkins; San Juan Hospital/Department of Pediatrics: Midnela Acevedo-Flores, Lourdes Angeli-Nieves, Vivian Olivera; SUNY Downstate Medical Center: Hermann Mendez, Ava Dennie, Susan Bewley; SUNY Stony Brook: Sharon Nachman, Margaret Oliver, Helen Rozelman; Tulane University Health Sciences Center: Russell Van Dyke, Karen Craig, Patricia Sirois; University of Alabama, Birmingham: Marilyn Crain, Newana Beatty, Dan Marullo; University of California, San Diego: Stephen Spector, Jean Manning, Sharon Nichols; University of Colorado Denver Health Sciences Center: Elizabeth McFarland, Emily Barr, Robin McEvoy; University of Florida/Jacksonville: Mobeen Rathore, Kristi Stowers, Ann Usitalo; University of Illinois, Chicago: Kenneth Rich, Delmyra Turpin, Renee Smith; University of Maryland, Baltimore: Douglas Watson, LaToya Stubbs, Rose Belanger; University of Medicine and Dentistry of New Jersey: Arry Dieudonne, Linda Bettica, Susan Adubato; University of Miami: Gwendolyn Scott, Claudia Florez, Elizabeth Willen; University of Southern California: Toinette Frederick, Mariam Davtyan, Maribel Mejia; and University of Puerto Rico Medical Center: Zoe Rodriguez, Ibet Heyer, Nydia Scalley Trifilio.
Disclaimer.The conclusions and opinions expressed in this article are those of the authors and do not necessarily reflect those of the National Institutes of Health or the US Department of Health and Human Services.
Financial support.This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cofunding from the National Institute of Allergy and Infectious Diseases, the National Institute on Drug Abuse, the National Institute of Mental Health, the National Institute of Deafness and Other Communication Disorders, the National Heart Lung and Blood Institute, the National Institute of Neurological Disorders and Stroke, and the National Institute on Alcohol Abuse and Alcoholism, through cooperative agreements with the Harvard University School of Public Health (U01 HD052102-04) and the Tulane University School of Medicine (U01 HD052104-01).
Potential conflicts of interest.All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed