Volume 50, Issue 3

Objectives A financial analysis is carried out to assess costs and benefits of providing cell-free DNA screening in Finland, using different strategies. Methods Three cell-free DNA screening strategies are considered: Primary, all women; Secondary, those with positive Combined test; and Contingent, the 10–30% with the highest Combined test risks. Three costs are estimated: additional cost for 10,000 pregnancies compared with the Combined test; ‘marginal’ cost of avoiding a Down syndrome birth which occurs in a pregnancy that would have been false-negative using the Combined test; and marginal cost of preventing the iatrogenic loss of a non-Down syndrome birth which occurs in a pregnancy that would have been false-positive. Results Primary cell-free DNA will require additional funds of €250,000. The marginal cost per Down syndrome birth avoided is considerably less than the lifetime medical and indirect cost; the marginal cost per unaffected iatrogenic fetal loss prevented is higher than one benefit measure but lower than another. If the ultrasound component of the Combined test is retained, as would be in Finland, the additional funds required rise to €992,000. Secondary cell-free DNA is cost-saving as is a Contingent strategy with 10% selected but whilst when 20–30% costs rise they are much less than for the Primary strategy and are cost-beneficial. Conclusions When considering the place of cell-free DNA screening it is important to make explicit the additional and marginal costs of different screening strategies and the associated benefits. Under most assumptions the balance is favorable for Contingent screening.

Objectives A systematic review and meta-analysis was conducted to quantitatively synthesize the current evidence on the association of prenatally diagnosed isolated single umbilical artery (iSUA) in singleton pregnancies with small for gestational age (SGA) neonates and other perinatal outcomes. Methods A search of PubMed/Medline, Scopus and The Cochrane Library was conducted, from inception to February 2021, in order to identify studies comparing the risk of SGA and other perinatal adverse outcomes in prenatally diagnosed iSUA singleton pregnancies vs. those with a 3-vessel cord (3VC). The quality of eligible studies was assessed according to the improved Newcastle–Ottawa Scale (NOS). The heterogeneity of results across the studies was tested using the I 2 test. Funnel plots and Egger’s test were used to assess the possibility of publication bias. Prospero RN: CRD42020182586. Results The electronic search identified 7,605 studies, of which 11 were selected, including three retrospective cohort and eight retrospective case control studies, overall reporting on 1,533 iSUA cases. The risk of delivering SGA neonates was increased in cases with iSUA (OR: 2.90; 95% CI: 2.02–4.18; p<0.00001; I 2 =71%). Similarly, iSUA was associated with an increased risk of pregnancy-induced hypertension (PIH) (OR: 2.23; 95% CI: 1.41–3.54; p<0.000; I 2 =1%), intrauterine death (IUD) (OR: 2.62; 95% CI: 1.43–4.79; p=0.002; I 2 =0%), preterm birth (PTB) (OR: 2.48; 95% CI: 1.73–3.56; p<0.00001; I 2 =56%), cesarean section (CS) (OR: 1.64; 95% CI: 1.11–2.41; p=0.01; I 2 =78%) and admission to neonatal intensive care unit (NICU) (OR: 2.28; 95% CI: 1.52–3.44; p<0.000001; I 2 =73%). Conclusions In prenatally diagnosed iSUA there is a higher risk of SGA, PIH, IUD, PTB, CS and NICU admission. These findings support the value of prenatal diagnosis of iSUA, which may subsequently intensify surveillance for the detection of specific pregnancy complications.

Objectives To compare perinatal outcomes in women with vs. without severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. Methods Perinatal outcomes in SARS-CoV-2 positive pregnant women who delivered at our institution between October 27th 2020 and January 31st 2021 were compared to SARS-CoV-2 negative pregnancies (contemporary controls) and historical 2019 controls matched by maternal age, pre-pregnancy body mass index and parity. Testing was performed based on symptoms or close contact at any time during pregnancy and as part of universal screening at hospital admission. Multivariable log-linear regression models were used adjusting for potential confounders (p < 0.05 statistically significant). Results One thousand three hundred seventeen women delivered at our institution during the study period. 1,124 (85%) tested negative and 193 (15%) positive for SARS-CoV-2. 189 (98%) were infected during third trimester. 19 (10%) were asymptomatic, 171 (89%) had mild to moderate coronavirus disease 2019 (COVID-19), and 3 (2%) were critically ill with one case of maternal death. There were no significant differences in preterm birth, small-for-gestational-age birth weight, congenital anomalies, operative delivery, intrapartum hypoxia, and perinatal mortality in SARS-CoV-2 positive pregnancies compared to contemporary reference group or historical controls from pre-COVID-19 period. Labor was more commonly induced in SARS-CoV-2 positive women compared to reference SARS-CoV-2 negative group (68 [35%] vs. 278 [25%], adjusted odds ratio 1.62; 95% confidence interval 1.14–2.28). Conclusions SARS-CoV-2 infection in pregnancy was not strongly associated with adverse perinatal outcomes. While the majority of SARS-CoV-2 positive women had no or mild/moderate symptoms, 2% were critically ill, with one case of maternal death.

Objectives This study aimed to examine how Turkish women were affected by the COVID-19 pandemic in the postpartum period in terms of postpartum support and anxiety variables. Methods The study was conducted with 130 women who gave birth during the COVID-19 period. Data were collected online using a personal information form, Postpartum Support Scale and Postpartum Specific Anxiety Scale. Descriptive statistics were used, with Kruskal–Wallis and regression analysis performed. Results All postpartum women stated that they were affected negatively by COVID-19. The most common issues described were feeling alone (25.51%), having economic difficulties (15.5%) and experiencing difficulties accessing health services (21.7%). The mean Postpartum Support Scale score was 102.43 ± 27.67, while the mean Postpartum Specific Anxiety Scale score was 188.07 ± 96.71. COVID-19 exposure level had a significant effect on postpartum-specific anxiety (p=0.00). The variables of having a problem during pregnancy (p=0.006), having a baby willingly (p=0.007), and partner’s educational status significantly predicted the degree of being affected by COVID-19 (p=0.025). The degree to which COVID-19 affected the women was determined by having problems during pregnancy, having a baby willingly, and a highly educated partner. Conclusions These variables may inform a resource to empower postpartum women during COVID-19. Healthcare systems need to be organised considering pandemic conditions to provide increased postpartum support, evaluate psychological health, deliver healthcare services, and consider pregnancy periods.

Objectives To introduce a first-line noninvasive antenatal management of maternal cytomegalovirus (CMV) primary infection based on ultrasound (US) and magnetic resonance imaging (MRI). Amniocentesis (AC) is used as a second-line tool in cases of abnormalities compatible with fetal CMV infection on US and/or MRI screening. Methods Between January 2011 and October 2018, pregnant women referred with a CMV primary infection on antibody screening were followed up by monthly US scans and a brain MRI at approximately 32 weeks. In cases with US and/or MRI abnormalities compatible with congenital CMV infection, AC was performed to confirm the diagnosis. Results Ninety pregnant women with a primary CMV infection were included (89 singleton and one twin pregnancy). The first-line screening by US and/or MRI was normal for 72 of 91 fetuses (79%). At birth, 19 of these 72 neonates (26%) had a positive urine sample for CMV but were asymptomatic. US and/or MRI abnormalities were identified in 19 fetuses (21%). AC confirmed a fetal CMV infection in 16 fetuses (84%); 12 pregnancies were terminated, and four were continued, with three symptomatic neonates at birth and one poor neurodevelopmental outcome at postnatal follow-up. Conclusions First-line noninvasive management of maternal CMV primary infection based on serial US scans and brain MRI can be offered to identify fetuses with severe symptomatic congenital CMV infection and reduce the number of ACs without compromising the fetal outcome.

Objectives We aimed to assess the performance of ultrasound (US) and magnetic resonance imaging (MRI) signs for antenatal detection of placenta accreta spectrum (PAS) disorders in women with placenta previa (placental edge ≤2 cm from the internal uterine orifice, ≥26 0/7 weeks’ gestation) with and without a history of previous Caesarean section. Methods Single center prospective observational study. US suspicion of PAS was raised in the presence of obliteration of the hypoechoic space between uterus and placenta, interruption of the hyperechoic uterine-bladder interface and/or turbulent placental lacunae on color Doppler. All MRI studies were blindly evaluated by a single operator. PAS was defined as clinically significant when histopathological diagnosis was associated with at least one of: intrauterine balloon placement, compressive uterine sutures, peripartum hysterectomy, uterine or hypogastric artery ligature, uterine artery embolization. Results A total of 39 women were included: 7/39 had clinically significant PAS. There were 6/18 cases of PAS with anterior placenta: hypoechoic space interruption and placental lacunae were the most sensitive sonographic signs (83%), while abnormal hyperechoic interface was the most specific (83%). On MRI, focal myometrial interruption and T2 intraplacental dark bands showed the best sensitivity (83%), bladder tenting had the best specificity (100%). 1/21 women with posterior placenta had PAS. There was substantial agreement between US and MRI in patients with anterior placenta ( κ =0.78). Conclusions US and MRI agreement in antenatal diagnosis of clinically significant PAS was maximal in high-risk women. Placental lacunae on ultrasound scan and T2 intraplacental hypointense bands on MRI should trigger the suspicion of PAS.

Objectives To assess the applicability of a standardized multidisciplinary protocol for managing placenta accreta spectrum (PAS) disorders and its impact on the outcomes. Methods We compared patients with PAS manage by a standardized multidisciplinary protocol (T2) to historic controls managed on a case-by-case basis by individual physicians between (T1). The primary outcome is composite maternal morbidity. Secondary outcomes were the rates of surgical complications, estimated blood loss, number of blood products transfused, intensive care unit admissions, ventilator use, and birth weight. Multivariate logistic analysis was used to identify independent predictors of composite maternal morbidity. Results During T1 and T2, we managed 39 and 36 patients with confirmed PAS, respectively. During T2, the protocol could be implemented in 21 cases (58%). Compared to T1, patients managed during T2 had 70% less composite maternal morbidity (95% CI: 0.11–0.82) and lower blood loss (median, 2,000 vs. 1,100 mL, p=0.008). Also, they were 68% less likely to require transfusion of blood products (95% CI: 0.12–0.81; p=0.01), including fewer units of packed red blood cells (median, 2 vs. 0, p=0.02). Management following the protocol was the only independent factor associated with lower composite maternal morbidity (OR: 0.22; 95% CI: 0.05–0.95; p=0.04). Selected maternal and neonatal outcomes were not different among 12 and 15 patients with suspected but unconfirmed PAS disorders managed during T1 and T2, respectively. Conclusions Most patients can be managed under a standardized multidisciplinary protocol for PAS disorders, leading to improved outcomes.

Objectives To compare between the effectiveness and safety of two different daily doses of vaginal progesterone (400 vs. 200 mg) in the prevention of preterm labor in twin pregnancy. Methods This is a prospective single-blinded randomized controlled trial conducted on 100 primi-gravida who had twin pregnancy and attended the antenatal clinic of a University hospital. They were equally and randomly allocated into two arms each containing 50 patients. Arm 1 received 400 mg and arm 2 received 200 mg vaginal progesterone daily at bed time starting from 14 weeks of pregnancy to 36 weeks. Transvaginal ultrasound was performed for assessment of the length of cervix at 14 and 22 weeks. Results Both arms of the current study were comparable regarding the cervical length at 14 and 22 weeks, mean gestational age at delivery, incidence of pre-term delivery, birth weight of the first twin. Second twin in addition to the average weight of both twins. No statistical significance differences between two arms regarding incidence of early neonatal death. NICU, mechanical ventilation, length of admission in NICU for the first twin. Second twin as well as both twins. Conclusions Vaginal progesterone treatment with different doses was tolerable, but wasn’t effective in the prevention of preterm labor in twin pregnancy.

Objectives To determine whether preimplantation genetic testing for aneuploidy (PGT-A) is associated with a reduced risk of abnormal conventional prenatal screening results in singleton pregnancies conceived using in vitro fertilization (IVF). Methods This was a retrospective cohort study of singleton IVF pregnancies conceived from a single tertiary care center between January 2014 and September 2019. Exclusion criteria included mosaic embryo transfers, vanishing twin pregnancies, and cycles with missing outcome data. Two cases of prenatally diagnosed aneuploidy that resulted in early voluntary terminations were also excluded. The primary outcome of abnormal first or second-trimester combined screening results was compared between two groups: pregnancy conceived after transfer of a euploid embryo by PGT-A vs. transfer of an untested embryo. Multivariable backwards-stepwise logistic regression with Firth method was used to adjust for potential confounders. Results Of the 419 pregnancies included, 208 (49.6%) were conceived after transfer of a euploid embryo by PGT-A, and 211 (50.4%) were conceived after transfer of an untested embryo. PGT-A was not associated with a lower likelihood of abnormal first-trimester (adjusted OR 1.64, 95% CI 0.82–3.39) or second-trimester screening results (adjusted OR 0.96, 95% CI 0.56–1.64). The incidences of cell-free DNA testing, fetal sonographic abnormalities, genetic counseling, and invasive prenatal diagnostic testing were similar between the two groups. Conclusions Our data suggest that PGT-A is not associated with a change in the likelihood of abnormal prenatal screening results or utilization of invasive prenatal diagnostic testing. Counseling this patient population regarding the importance of prenatal screening and prenatal diagnostic testing, where appropriate, remains essential.

Objectives An association between fetal growth restriction (FGR) and increased predisposition to cardiovascular disease (CVD) is suggested. The aim of this study was to evaluate subclinical signs of fetal cardiac remodeling in late-onset small-for-gestational-age (SGA) and growth-restricted fetuses using two-dimensional speckle tracking echocardiography (2D-STE). Methods This is a prospective cohort study, including 117 late-onset (≥32 weeks) SGA (birthweight≤10th centile) fetuses and 102 gestational age matched controls. A subgroup analysis was performed: FGR was defined based on either BW (<third centile) or the presence of feto-maternal Doppler abnormalities, the remaining cases were called SGA centile 3–10. 2D-STE based myocardial deformation analyses included longitudinal peak systolic strain (LPSS) and strain rate (LSR) of the global left (LV) or right (RV) ventricle and its individual segments. Results Global and segmental LPSS and LSR values showed no significant difference between late-onset SGA and control fetuses for both ventricles (p>0.05). Regarding global and segmental LPSS and LSR values of LV/RV, subgroup analysis revealed no significant difference between the FGR (n=81), SGA centile 3–10 (n=36) and control group. Conclusions A mild degree of placental dysfunction seems not to influence myocardial deformation properties measured by 2D-STE.

Objectives To assess fetal cardiac structures using HDlive Flow Silhouette with spatiotemporal image correlation (STIC) at 12 to 14 + 6 weeks of gestation, and verify the feasibility of obtaining five cardiac views in the late first and early second trimesters of pregnancy. The fetal cardiac shape and the aspect of the descending aorta were also evaluated. Methods Eighty normal fetuses at 12 to 14 + 6 weeks of gestation were studied using trans-abdominal HDlive Flow Silhouette with STIC to assess the feasibility of five fetal cardiac views (frontal, spatial three-vessel, panoramic, posterior, and right lateral views). Target structures in each cardiac view were evaluated. ‘Good’ was assigned when all structures were noted, ‘Fair’ when only one structure was missed, and ‘Poor’ when two and more structures could not be detected. Frequencies of an elongated heart and those of a tortuous descending aorta were counted. Results Forty-nine fetuses were effectively included in the analysis. Success rates of ‘Good’ and ‘Fair’ were significantly higher with spatial three-vessel (p<0.01) and panoramic views (p<0.05). Frequencies of “Elongated heart”, “Elongated left ventricle”, and “Spherical heart” were 12.2, 6.2, and 81.6%, respectively. Frequencies of “Tortuous descending aorta” and “Straight descending aorta” were 12.2 and 87.8%, respectively. Conclusions The feasibility of obtaining fetal five cardiac views using HDlive Flow Silhouette with STIC is good, and this technique provides useful information for evaluating fetal cardiac structures in the late first and early second trimesters of pregnancy.

Objectives To analyze umbilical artery (UA) Doppler velocimetry and its possible role in placenta-mediated fetal growth restriction (FGR) in second- and third-trimester fetuses with trisomy 18 and 13. Methods UA pulsatility index (PI) and half-peak systolic velocity deceleration time (hPSV-DT) were measured in fetuses with trisomy 18 and 13. Correlation with gestational age, birthweight, and perinatal outcome was analyzed. Results A total of 80 measurements were taken from 33 fetuses with trisomy 18 and 19 with trisomy 13. Overall, there was a high prevalence of abnormal UA Doppler velocimetry. In fetuses with trisomy 18, 54% (27/50) of the UA PI values and 58% (29/50) of the UA hPSV-DT values were abnormal. In fetuses with trisomy 13, 80% (24/30) of the UA PI values and 87% (26/30) of the UA hPSV-DT values were abnormal. The prevalence of abnormal UA Doppler velocimetry increased with gestational age in both types of aneuploidy. However, this trend was only significant for trisomy 13 (p<0.05). All fetuses with trisomy 18 and 86% of fetuses with trisomy 13 were classified at birth as FGR. There were no perinatal survivors in this series. Conclusions A high prevalence of abnormal UA Doppler velocimetry was found in second- and third-trimester fetuses with trisomy 18 and 13, which further increased with gestational age. These results may well correlate with alterations described previously in the placenta, suggesting placental insufficiency has an important role in the development of FGR in these autosomal aneuploid fetuses.

Objectives Pulmonary hypertension (PH) is a complication of bronchopulmonary dysplasia (BPD) and associated with increased mortality and morbidity. Our aim was to identify, in infants with BPD, the effect of PH on health-care utilisation and health related cost of care. Methods An electronic data recording system was used to identify infants ≤32 weeks of gestation who developed BPD. PH was classified as early (≤28 days after birth) or late (>28 days after birth). Results In the study period, 182 infants developed BPD; 22 (12.1%) developed late PH. Development of late PH was associated with a lower gestational age [24.6 (23.9–26.9) weeks, p=0.001] and a greater need for positive pressure ventilation on day 28 after birth (100%) compared to infants without late PH (51.9%) (odds ratio (OR) 19.5, 95% CI: 2.6–148), p<0.001. Late PH was associated with increased mortality (36.4%) compared those who did not develop late PH (1.9%) after adjusting for gestational age and ventilation duration (OR: 26.9, 95% CI: 3.8–189.4), p<0.001. In infants who survived to discharge, late PH development was associated with a prolonged duration of stay [147 (118–189) days] compared to the infants that did not develop late PH [109 (85–149) days] (p=0.03 after adjusting for gestational age). Infants who had late PH had a higher cost of stay compared to infants with BPD who did not develop late PH (median £113,494 vs. £78,677, p=0.016 after adjusting for gestational age). Conclusions Development of late PH was associated with increased mortality, a prolonged duration of stay and higher healthcare cost.

Objectives We evaluated inpatient management, transition to home, breastfeeding, growth, and clinical outcomes of infants born to mothers diagnosed with SARS-CoV-2 infection in pregnancy and followed in a Federally Qualified Health Center (FQHC), that serves a diverse and low-income patient population, from birth through 6 months of life. Methods Infants born between 4/3/20 and 7/26/20 at Prentice Women’s Hospital with mothers who received prenatal care at Erie Family Health Center (Erie), the second largest FQHC in Illinois, and had confirmed SARS-CoV-2 during pregnancy were included. Data were abstracted from delivery hospital admission and outpatient follow-up appointments between 4/8/20 and 2/4/21. Results Thirty-three infants met inclusion criteria. Average gestational age was 38.9 weeks (IQR 37.6–40.4), 3 (10%) were premature and 5 (15%) required NICU admission. Nearly all (97%) mothers expressed intent to breastfeed. Outpatient follow-up rates were similar to historical cohorts and 82% (23/28) of infants were vaccination compliant. Growth parameters showed normal distributions at all time points. At 6 months, any and exclusive breast milk feeding rates were lower compared to historic cohorts (18 vs. 36%, p<0.05, 0 vs. 21%, p<0.01). Three infants (10%) received development-related referrals, one carried an underlying genetic diagnosis. Outpatient visits were predominantly face-to-face with telemedicine use comprising only 6% of visits (11/182). Conclusions Longitudinal follow-up of 33 publicly insured infants born to mothers with SARS-CoV-2 infection in pregnancy followed in an FQHC showed high rates of follow-up and vaccination compliance, normal growth patterns and reassuring clinical status, and lower than expected rates of breastfeeding.

Objectives Although neonates with moderate to severe hypoxic ischemic encephalopathy (HIE) receive therapeutic hypothermia (TH), 40–50% die or have significant neurological disability. The aim of this study is to analyse the association of placental pathology and neurodevelopmental outcome in cooled neonates with HIE at 18–24 months of age. Methods Retrospective analysis of prospectively collected data on 120 neonates registered in the Swiss National Asphyxia and Cooling Register born between 2007 and 2017. This descriptive study examines the frequency and range of pathologic findings in placentas of neonates with HIE. Placenta pathology was available of 69/120 neonates, whose results are summarized as placental findings. As neonates with HIE staged Sarnat score 1 (21/69) did not routinely undergo follow-up assessments and of six neonates staged Sarnat Score 2/3 no follow-up assessments were available, 42/48 (88%) neonates remain to assess the association between placental findings and outcome. Results Of the 42/48 (88%) neonates with available follow up 29% (12/42) neonates died. Major placenta abnormalities occurred in 48% (20/42). Major placenta abnormality was neither associated with outcome at 18–24 months of age (OR 1.75 [95% CI 0.50–6.36, p=0.381]), nor with death by 2 years of age (OR 1.96 [95% CI 0.53–7.78, p=0.320]). Conclusions In this study cohort there could not be shown an association between the placenta findings and the neurodevelopmental outcome at 18–24 months of age.

Objectives The current study aims to assess different high-frequency (HF) band power calculations based on different frequency bandwidth values, and compare them with the time domain the root mean square of successive RR differences (RMSSD) value in preterm infants. Methods At week 32, electrocardiogram (ECG) and breathing rate (BR) were recorded for 24 h on 30 preterm infants born between 28 and 32 weeks. The recording held in the neonatal intensive care unit without any interruption of routine. Results The median 24 h BR was 40–78 breaths per minute. The RMSSD was highly and positively correlated with frequency bands that were based on each preterms BR range, or on a constant frequency with band limits of 0.4–2 Hz. Conclusions At week 32, HF band Hz limits should be calculated based on each child’s breathing rate, generally between 0.4 and 2 Hz.

Objectives The aim of this study is to evaluate the diagnostic ability of multiplex real-time polymerase chain reaction (PCR) in very preterm infants assessed for risk of early onset neonatal sepsis (EOS). Methods Prospective observational cohort study. Blood samples of preterm neonates ≤32 weeks of gestation were evaluated by commercial multiplex real-time PCR within 2 h after delivery. The definition of EOS was based on positive blood culture and clinical signs of infection or negative blood culture, clinical signs of infection and abnormal neonatal blood count and serum biomarkers. Results Among 82 subjects analyzed in the study, 15 had clinical or confirmed EOS. PCR was positive in four of these infants (including the only one with a positive blood culture), as well as in 15 of the 67 infants without sepsis (sensitivity 27%, specificity 78%). Out of 19 PCR positive subjects, Escherichia coli was detected in 12 infants (63%). Statistically significant association was found between vaginal E. coli colonization of the mother and E. coli PCR positivity of the neonate (p=0.001). No relationship was found between neonatal E. coli swab results and assessment findings of bacterial DNA in neonatal blood stream. Conclusions Multiplex real-time PCR had insufficient diagnostic capability for EOS in high risk very preterm infants. The study revealed no significant association between PCR results and the diagnosis of clinical EOS. Correlation between maternal vaginal swab results and positive PCR in the newborn needs further investigation to fully understand the role of bacterial DNA analysis in preterm infants.

Objectives We designed an electronic query that was distributed to healthcare workers in order to evaluate their attitudes towards COVID-19 vaccination of women that wish to conceive and those that are pregnant. Methods An electronic survey was designed and distributed through a network of 5.654 e-mails. The query was uploaded in Google Forms. Results Overall, 1,226 answers were retrieved. Ninety percent of respondents replied that they prescribe COVID-19 vaccination in women that wish to conceive, however, the rate falls to 80% for those that have schedules for imminent artificial reproductive techniques. Approximately 1 in 2 physicians (obstetricians and pediatricians) suggests that a delay in assisted reproduction following immunization is not recommended. At least 13% of women of reproductive age and 30% of women wishing to conceive with ART declined to do so according to respondents, report. Only 75% of respondents favored immunization during pregnancy and of those 48% responded that vaccination could be done in all trimesters of pregnancy. Approximately 40% of women decline; however, to receive vaccination. Conclusions Despite the directions provided by international organizations; physician and patient hesitancy towards vaccination during pregnancy (and particularly the first trimester) remains an issue that deserves further attention.