ATLANTIC DIP, Changing the Landscape for Better Maternal and Infant Health: The 2024 Norbert Freinkel Award Lecture Free

In 2005, I became lead for the Diabetes in Pregnancy service. In preparation for leading the service, I embarked on a fact-finding mission. Overall, I found a lack of data on key indicators of care and pregnancy outcomes in women with pregestational diabetes mellitus (PGDM). Information on GDM prevalence or outcomes was unavailable, and a wide variety of GDM screening practices, protocols, and guidelines were in place across the region. To address these deficiencies, I secured a program grant from the Health Research Board (HRB) to establish the ATLANTIC DIP network. The geographical area of ATLANTIC DIP covers one-quarter the landmass of Ireland, measuring 350 × 250 km, with a population of 750,000 mixed urban and rural dwellers and 10,000 deliveries per year across five antenatal centers (Fig. 1). The 5-year objectives of ATLANTIC DIP were to 1) establish a clinical and research network, 2) develop and implement an electronic data collection tool, 3) define the maternal and neonatal outcomes of pregnancies complicated by PGDM, 4) establish GDM prevalence and outcomes, and 5) establish the prevalence for dysglycemia after having GDM. We chose DIAMOND (Hicom UK), an electronic diabetes patient information system, as our collection tool, and we designed the prepregnancy, pregnancy, and postpregnancy fields to ensure collection of all appropriate data.

We systematically examined regional pregnancy outcomes in women with PGDM using the novel data collection tool (1). Women with PGDM experienced more adverse pregnancy outcomes, with congenital malformation rate two times, perinatal mortality rate three times, and stillbirth rate five times those of the background population (Fig. 2). Women were not well prepared for pregnancy; only 28% had received prepregnancy care (PPC), 43% had received prepregnancy folic acid, and 50% had an HbA_1c <7% (53 mmol/mol) (Fig. 2). With this information and a review of the literature (2–4), we concluded that pregnancy outcomes could be improved by a regional, protocol-driven PPC program. We also implemented a number of low-cost interventions, including prospective collection of high-quality data, education to improve knowledge and skills, and introduction of a PPC leaflet, a patient guide book to explain goals, and a clinician guide to facilitate clinical decision-making (5). Audit of regional data is a powerful way to change culture and practice, and low-cost interventions can be associated with big gains in practice and outcomes. This can be achieved in almost any jurisdiction.

With further funding (from HRB), we designed, implemented, and analyzed a single-arm regional PPC intervention program. The design removed barriers to attendance by providing the PPC intervention at all antenatal centers. A central DIP coordinator facilitated attendance and data collection. Referrals were accepted from specialist providers, primary care clinicians, and self-referral by women. A standardized form based on NICE (U.K. National Institute for Health and Care Excellence) guidance was used (6). The intervention included patient education, medication review, assessment and treatment of complications, assessment of thyroid status, administration of folic acid 5 mg/day, intensive glucose monitoring, and a target HbA_1c <6.5% (48 mmol/mol) (7). The endocrinologist, diabetes nurse specialist, and dietitian worked together to deliver the program. The program was shown to be both clinically effective and cost-effective (8). Attendees were less likely to smoke (P = 0.03), had a high uptake of folic acid (P < 0.001), eliminated teratogenic medications (P = 0.008), and had a lower mean HbA_1c in the first trimester (P = 0.004) (Fig. 2). A lower HbA_1c was maintained throughout pregnancy for attendees (P < 0.001). This translated into improved neonatal outcomes with reduction in serious adverse neonatal events (P = 0.007) and reduction in congenital malformations (P = 0.04), and the need for care in the neonatal intensive care unit (NICU) was reduced by one-third (P = 0.002) (Fig. 2). There was a savings of EUR 2,577 for each woman attending PPC, driven by a reduction in the cost of treatment of congenital malformations and cost of NICU care (8). In view of these results, the PPC program became part of routine clinical care and continues in the region. PPC is a powerful intervention, is recommended by international guidelines (6), and is available in many services globally. Regardless of location, the components of a PPC intervention are uniform.

Although the results of the PPC intervention program demonstrated significant benefits, we reviewed the totality of our regional interventions to see if there was a lasting impact on our PGDM population. Publications demonstrate that the interventions have had a significant and sustained effect, with reductions in stillbirth, perinatal mortality, and congenital malformation rates to levels seen for the background population. These improved and sustained outcomes were driven by an increase in the uptake of folic acid and a lowering of HbA_1c at conception, with maintenance of this lower HbA_1c throughout pregnancy (9–11). In 2015, we commenced a national audit process to assess perinatal outcomes of pregnancies complicated by PGDM nationally in 19 antenatal centers with voluntary participation (12,13). These national audits have directed the clinical community toward several areas for improvement.

Good retinal care is important for women with PGDM (type 1 and type 2) during pregnancy. Retinopathy may progress in those with rapid improvement of glycemia and in those with hypertension or preeclampsia. It is recommended that retinal screening occur on at least two occasions during pregnancy. We examined the pregnancy retinal screening process using a population-based cohort study (14). We found that 40% of the cohort had no or inadequate screening. In the remaining 60% who received appropriate screening, retinopathy progression occurred in 26%. Our findings indicated a poor process of care and a missed opportunity where an intervention would likely be successful. The results of this cohort study propelled us toward a solution nationally. As of 2023, all pregnant women with diabetes now receive two additional evaluations as part of the national diabetes retinal screening program, with >90% uptake (15). In many countries retinal screening occurs independently of diabetes antenatal care, often in community settings. Information on the screening visit is often not available to the clinical team providing diabetes and antenatal care. For implementing a national program, applying the principles and governance of population screening is the solution. It provides a central source of information and accessible evidence of retinopathy status. Such information is important in decision-making throughout the antenatal journey and delivery planning for women with PGDM.

In 2008, we commenced studies to establish GDM prevalence and outcomes. In a population-level cohort study, we invited 12,487 consecutive pregnant women at five antenatal centers in the ATLANTIC DIP region to attend universal GDM screening with a 75-g oral glucose tolerance test (OGTT) at 24–28 gestational weeks irrespective of risk factors. Three-quarters (n = 9,242) consented, while 25.9% (n = 3,237) declined screening. Of those consenting, 44% attended (n = 5,500) testing while 30% (n = 3,742) did not attend. The prevalence of GDM based on International Association of the Diabetes and Pregnancy Study Groups criteria was 12.4% (16) (Fig. 3). Women with GDM displayed higher rates of pregnancy-induced hypertension (PIH) (86 [13.8%] vs. 332 [7.5%]; P < 0.0001) and preeclampsia (39 [6.3%] vs. 176 [4%]; P = 0.007) and were more likely to require a cesarean delivery (246 [37.2%] vs. 1,165 [24.9%]; P < 0.0001) compared with women without GDM (Fig. 3 and Table 1). Infants of mothers with GDM were delivered earlier and were more likely to experience preterm birth (47 [7.1%] vs. 223 [4.8%]; P = 0.002) and macrosomia (33 [23.9%] vs. 156 [17%]; P < 0.05), to be born large for gestational age (LGA) (149 [22.6%] vs. 751 [16.2%]; P < 0.0001), and to require NICU care (170 [26%] vs. 419 [9.1%]; P < 0.0001) compared with infants of mothers without GDM (Fig. 3 and Table 1) (16). Using a geographical information system, we geocoded participant addresses and identified that for every 10 km required to travel for screening, the probability of attending was reduced by 2%; thus, if a woman lived 100 km away from the screening site she was 20% less likely to attend (17). With this novel information, we designed a randomized trial of GDM screening sites. We hypothesized that screening in primary care compared with hospital care would have a lower nonattendance rate. Seven hundred eighty-one women were randomized, half to each screening site. In the group assigned to hospital screening there was a lower nonattendance rate compared with those assigned to primary care (8.5% vs. 14.8%) (18). In addition, on a cost-effectiveness analysis, screening for GDM in the hospital was most cost-effective (19). The results also demonstrated that implementation of universal screening would be cost-effective (19).

We revisited the original universally screened cohort (16) to examine if women with GDM but without risk factors had better pregnancy outcomes than those with GDM and risk factors. We applied current risk factor screening guidelines to this universally screened cohort to examine the performance of these guidelines in an Irish population. Pregnancy outcomes of women who had no risk factors but had GDM were more adverse, with higher rates of hypertensive disorders (13% vs. 7%; P = 0.029) and cesarean deliveries (38% vs. 24%; P = 0.0001), compared with women with normal glucose tolerance (NGT) (20). When NICE, Irish, and American Diabetes Association guidelines were applied, 20% (n = 120), 16% (n = 101), and 5% (n = 31), respectively, of women with GDM without risk factors would have gone undiagnosed (20). This highlighted the drawbacks of risk factor–based screening in the Irish setting and the missed opportunities for health gain. At this juncture, we implemented a robust regional GDM screening program using World Health Organization 2013 criteria (21), but as our analysis provided a strong argument for universal screening, we have continued to advocate for this.

Focusing on the ongoing disease burden after a diagnosis of GDM, women from the universally screened cohort (GDM and NGT) were invited for assessment in the first postpartum year. Two hundred seventy women with GDM and 388 with NGT participated. Forty-two (16.6%) women with GDM compared with three (0.8%) with NGT had persistent dysglycemia (22), and breast feeding was shown to be protective against dysglycemia. We reanalyzed women at 5 years after the index pregnancy. By this time, 25.9% of women with GDM compared with 3.6% of women with NGT had diabetes or prediabetes (23). In view of the high burden of ongoing metabolic abnormalities, we looked to simplify ongoing community follow-up. We demonstrated that 90% of women with abnormal glucose tolerance are identified using cutoffs of ≥39 mmol/mol (5.7%) for HbA_1c and ≥5.6 mmol/L (100 mg/dL) for fasting glucose, with a 66% reduction in the need for a 75-g OGTT (24). These cutoffs are now recommended to primary care practitioners for the long-term follow-up of women with GDM.

GDM describes women who are first recognized with a glucose abnormality in pregnancy, but GDM is heterogeneous. We analyzed biobank samples from our universally screened cohort to establish the prevalence of glucokinase monogenic diabetes (maturity-onset diabetes of the young [MODY]) in a regional pregnant population. The prevalence of glucokinase MODY was 0.12% in the total population and 0.93% in those diagnosed with GDM. Important maternal characteristics to identify GDM with glucokinase MODY were fasting blood glucose (FBG), which was significantly higher (6.4 mmol/L [115 mg/dL] vs. 5.2 mmol/L [94 mg/dL]; P < 0.0001), and maternal BMI (23.0 vs. 31.2 kg/m²; P < 0.0001), which was significantly lower than levels in women with GDM without this mutation (Table 2). We demonstrated that glucokinase MODY prevalence increased further to 1.62% and 2.56% when the FBG was >5.1 mmol/L (92 mg/dL) and >5.5 mmol/L (99 mg/dL), respectively (25). Based on these findings, we offer MODY glucokinase testing to mothers with a BMI ≤25 and FBG ≥5.5 mmol/L (99 mg/dL). These findings have international relevance and can be globally applied.

Women with diabetes during pregnancy are at higher risk of adverse perinatal outcomes. Excessive gestational weight gain (GWG) has emerged as a risk factor for perinatal complications (26), and the Institute of Medicine (IOM) published recommendations for appropriate GWG (27). In 2014, it was unclear whether excessive GWG conferred additional risk to women with diabetes and whether the IOM recommendations were applicable to these women. We designed an observational study conducted across five antenatal centers to examine whether excessive GWG in pregnancies with diabetes was associated with higher adverse perinatal outcomes. Excessive GWG was noted in 59% of our participants. Excessive GWG resulted in higher odds for LGA (adjusted odds ratio [aOR] 2.01, 95% CI 1.24–3.25 in GDM; aOR 3.97, CI 1.85–8.53 in PGDM) and macrosomia (aOR 2.17, CI 1.32–3.55 in GDM; aOR 3.58, CI 1.77–7.24 in PGDM). Excessive GWG was associated with increased odds for PIH (aOR 1.72, CI 1.04–2.85), and treatment with insulin increased further the odds for LGA (aOR 2.80, CI 1.23–6.38) and macrosomia (aOR 5.63, CI 2.16–14.69) in GDM (Table 3). With these findings, we advocated the use of IOM recommendations in women with GDM or PGDM and recommended targeted management of GWG particularly in those who require insulin. We implemented this pragmatically by insisting that maternal weight be captured at each clinic visit and the woman given a weekly weight target according to BMI. The weekly weight target was discussed alongside blood glucose targets and other parameters at each clinic visit. Future research should be directed toward preventing or reversing excessive GWG and its potential benefit on adverse perinatal outcomes (28). In view of our findings on excess GWG, we reviewed our GDM cohort treated by lifestyle intervention alone and found this intervention to be highly effective, with a reduction in LGA and macrosomia in BMI-matched women with NGT. We also found that the offspring were 21% less likely to have an adverse outcome (29). We then reviewed data of women with GDM who were treated with insulin. In addition to excessive GWG, we found an increase in the rate of cesarean delivery and a greater need for infants to receive NICU care (30). With the findings from this entire body of work, we examined other treatment options in the literature and designed and delivered the trial of early metformin in addition to usual care in the reduction of GDM effects (EMERGE) (31).

EMERGE was a phase III parallel, randomized, double-blind, placebo-controlled, superiority trial examining the effectiveness of early metformin in addition to usual care in the reduction of GDM effects (32). We designed EMERGE to see if metformin added to usual care at the time of a GDM diagnosis reduced the need for insulin or FBG ≥5.1 mmol/L (92 mg/dL) at week 32 or 38. We chose this composite primary outcome because it measures metformin failure in two discreet ways and also accounts for clinician bias in the decision to use insulin. Secondary outcomes included infant birth weight, length, and head circumference, with derived measures for LGA (>90th percentile), small for gestational age (SGA) (<10th percentile), macrosomia (>4-kg infant birth weight), low birth weight (<2.5-kg infant birth weight), and proportion of infants with neonatal morbidities, including need for NICU care, respiratory distress requiring respiratory support, jaundice requiring phototherapy, major congenital anomalies, Apgar score <7 at 5 min, and neonatal hypoglycemia <2.6 mmol/L (47 mg/dL) on one or more occasions within 60 min postdelivery. Secondary maternal outcomes included time to insulin initiation, insulin dose required, development of PIH or preeclampsia, antepartum and postpartum hemorrhage, mode and time of delivery, number of preterm births (<37 weeks’ gestation), GWG, self-reported capillary glucose, and self-reported treatment satisfaction.

Over a 5-year period (spanning the COVID-19 pandemic), 535 pregnancies were randomized, 268 and 267 to metformin and placebo, respectively. Information on 526 (98%) participants and their infants was available for the primary outcome and neonatal pregnancy outcomes. The primary outcome did not differ between groups. However, participants assigned to metformin were 25% less likely to require insulin (P = 0.004) and had a significantly lower FBG at gestational weeks 32 (P = 0.03) and 38 (P < 0.001) (Fig. 4A and B and Table 4). Participants randomized to metformin had less GWG (P = 0.003) and were more satisfied with their treatment (P = 0.04) (Table 4). The mean gestational age at delivery was similar between groups (39.1 weeks), with no difference in number of preterm births between groups. The mean birth weight of infants was lower in participants randomized to metformin, with a mean difference of –113 g (P = 0.01) and a lower proportion of infants displaying macrosomia (7.6% vs. 14.8%) (P = 0.02) or LGA (6.5% vs. 14.9% (P = 0.003) (Table 4). Numerically there were more infants born SGA in the metformin group, but this was not statistically significant (P = 0.13). A secondary analyses of SGA infants demonstrated rates of SGA similar to those of other trials, no significant increase in the odds of SGA with metformin, and no statistically significant difference in maternal characteristics between those who did and did not have SGA infants. Exposure to metformin was not associated with a more adverse SGA phenotype. The presence of hypertensive disorders of pregnancy was strongly associated with SGA independent of metformin exposure (33).

While the results of EMERGE show benefits to mothers and infants with no worsening of phenotype in SGA infants exposed to metformin, controversy and debate continues regarding the use of metformin in GDM. An emerging issue to consider is the possible impact of metformin on testicular tissue. In EMERGE, we found no excess genitourinary anomalies in metformin-exposed male infants. The trial was not powered to show a difference in such abnormalities, however, as a far larger sample size is needed to detect between-group differences. It is important that long-term follow-up of the mother-infant pairs is conducted to examine the medium- and long-term benefits or harms of metformin exposure, and this has now started for the EMERGE mothers and their children. Ongoing assessment of offspring anthropometrics and level of adiposity in addition to neurocognitive reporting are essential.

Since the birth of ATLANTIC DIP, we continue to improve the process of care with improved pregnancy outcomes. We have moved from a regional to a national collaborative effort, working with the national diabetes clinical care program of the Department of Health. Through a multistakeholder effort, we have developed a national model of care, which was launched by the Minister for Health in 2024 (34). This document will facilitate equitable national delivery of care using agreed national guidelines and processes and facilitate future funding. In 2022, retinal screening in pregnancy was incorporated into the national retinal screening program (15), and all women with PGDM are offered retinal screening on two occasions. If retinopathy requiring treatment is identified, women receive rapid assessment in a retinal clinic. The initial review of the program shows an uptake of >90% for screening (15). The national audit process of pregnancy outcomes initiated by ATLANTIC DIP has now been incorporated centrally in the National Office for Clinical Audit. This ensures its security in perpetuity, encouraging full participation of all antenatal centers nationally. Working with the National Office for Clinical Audit, it is our ambition to also develop a national audit tool for GDM. Since January 2023 and supported by evidence from ATLANTIC DIP, additional funds have been allocated by the Minister for Health to allow women with GDM to enter the National Disease Prevention Program, administered in the community setting and delivered by primary care physicians. It targets the prevention of type 2 diabetes, obesity, and cardiovascular disease in this population. Women are offered an annual screen with targeted interventions (35). Finally, based on our analysis of risk factor screening for GDM, an application to the National Screening Advisory Committee of the Department of Health to move to universal screening has received a favorable review and has moved to the next stage of a full health technology assessment (36).

In 2016, we established the international group INternational collaboration for Studies in Pregnancy and Diabetes (INSPIRE). Contribution to this group was global, with participation from researchers across Europe, Australia, New Zealand, and North and South America. The focus of the group was to develop core outcome sets (COS) for studies and trials in diabetes in pregnancy that could be used globally. COS use a well-known and referenced international methodology. They contain an agreed minimum set of outcomes to be measured and reported in all studies in a specific area, with the objective of standardizing outcome reporting. COS may minimize research waste by identifying outcomes important to key stakeholders, allowing for improved evidence synthesis, and facilitating translation of research findings into clinical practice (37). Some of the work from INSPIRE was completed on a virtual platform during the COVID-19 pandemic. In 2017, a COS of what should be collected to examine the effectiveness of PPC in those with PGDM was completed (38), and 17 outcomes were selected. We advocate that all trials and studies in PPC use this COS to improve transparency and increase our ability to compare and combine future global studies with greater ease. In 2019, a COS for the follow-up of women with GDM treated with insulin and/or oral glucose-lowering agents was completed. Nine outcomes were included in the final COS (39). Use of this COS will bring consistency and uniformity to outcome selection and reporting in clinical trials in GDM across the globe. In 2020, a COS for studies of GDM prevention and treatment was completed (40), with 14 outcomes selected. Future GDM prevention and treatment trials should refer to this COS to ensure that comparison and combination of trial results can occur. In 2021, a COS for the treatment of pregnant women with PGDM was completed (41), and 19 outcomes were included. This COS will enable better comparison between trials to produce robust evidence synthesis, improve trial reporting, and optimize research efficiency. In 2022, we examined the quality of patient-reported outcome (PRO) reporting in trials of DIP using a systematic review (42). We identified 7,122 citations with 35 articles included for review. Only 17% of randomized controlled trials included a PRO as a primary or secondary outcome. Out of a maximum score of 100, the median score was 46, indicating suboptimal reporting of PROs. Researchers should be mindful of the importance of PRO inclusion and reporting and include reliable PROs in trials. These internationally developed COS are gaining traction, as funding bodies and journal editors encourage their use. This body of international work highlights the importance of collaboration across continents. It also illustrates how the development of impactful research outputs is not costly.

Through the organizational structure of ATLANTIC DIP, we have contributed to international trials, ensuring women in Ireland were included in the investigation of innovative interventions, medicines, and technologies. These trials have all changed international practice. Vitamin D and Lifestyle Intervention for GDM Prevention (DALI) was an exciting program funded by the European Union and involved 11 trial centers in nine European countries. The trial showed that a combination of healthy eating and physical activity was the most effective strategy for limiting GWG but did not reduce fasting glycemia (43). In DALI, we demonstrated the heterogeneity of populations in Europe with reference to obesity, the prevalence of GDM in women with obesity across European nations, including the barriers to screening, and finally the prevalence of GDM across trimesters in those randomized to DALI (44–46). DALI has contributed >30 publications to the international literature. ATLANTIC DIP was selected to be a recruitment site for the Continuous Glucose Monitoring in Women With Type 1 Diabetes in Pregnancy Trial (CONCEPTT), which has been instrumental in changing clinical practice in PGDM (47). The investigators showed that the use of CGM during pregnancy in patients with type 1 diabetes was associated with improved neonatal outcomes, likely attributable to reduced exposure to maternal hyperglycemia. This study was the first to indicate potential for improvements in nonglycemic health outcomes from CGM use and advocated the use of CGM to all pregnant women with type 1 diabetes. Many guidelines (NICE and American Diabetes Association) now advocate for the use of CGM in women with type 1 diabetes, and in many countries this is now the standard of care (6,48). We participated in EVOLVE (An International Noninterventional Cohort Study to Evaluate the Safety of Treatment With Insulin Detemir in Pregnant Women With Diabetes Mellitus), a European postmarketing surveillance study of the use of long-acting insulin detemir in PGDM pregnancy. It examined the risk of major congenital malformations or perinatal or neonatal death with insulin detemir compared with other basal insulins in pregnant women with PGDM (49). Insulin detemir was associated with a risk profile similar to that of other basal insulins and can be used safely in pregnancy. Finally, we participated in the EXPECT (Research Study Comparing Insulin Degludec to Insulin Detemir, Together With Insulin Aspart, in Pregnant Women With Type 1 Diabetes) trial, an open-label, multinational, randomized, controlled, noninferiority trial of insulin degludec versus insulin detemir, both in combination with insulin aspart in the treatment of pregnant women with type 1 diabetes (50). A total of 225 women with type 1 diabetes were randomly assigned to degludec (n = 111) or detemir (n = 114), and degludec was found to be noninferior to detemir. There were no safety concerns, and degludec can now be used safely in pregnancy. Large multinational trials, such as those described above, are influential in changing clinical practice, updating guidelines, and influencing national and international policy. Clinical sites should be in a state of readiness to pivot and contribute to such trial efforts.

ATLANTIC DIP has been instrumental in changing the landscape for pregnant women with diabetes in Ireland. It has established evidence for prevalence and outcomes of pregnancies complicated by PGDM and GDM. It has identified shortcomings in the process of delivery of care and has developed solutions that are now firmly embedded. ATLANTIC DIP data have also contributed to the international literature and informed policy and guidelines. The EMERGE trial has provided robust evidence for the use of metformin in GDM management. Over the next decade, we will focus on three areas: 1) follow-up of EMERGE mothers and their children to document the impact of metformin use during pregnancy on the ongoing health status of both, 2) establish a national audit for pregnancy outcomes in women with GDM, and 3) deliver an RCT of the clinical and economic impacts of real-time CGM in women with GDM. Above all, we will continue to advocate for pregnant women with diabetes, striving for even better outcomes and access to innovative medicines and technologies. Healthy mothers are more likely to have healthy babies and promote healthy family units. We are in a privileged position to be able to contribute to this cycle of events.

Acknowledgments. I would like to thank all the women and their babies who participated in the ATLANTIC DIP program and the EMERGE trial who have contributed to this body of knowledge. In addition, I thank the University of Galway and Saolta Hospital group leadership teams who enabled the research and the HRB for ongoing grant support. Finally, I thank the clinical and research staff who facilitated the delivery of the program and my medical colleagues and those in training, all of whom have made significant contributions.

Funding. The EMERGE trial was funded by the HRB (31–33). Metformin and placebo were received through a charitable gift from Merck, and the blood glucose meters and strips were received through a charitable donation from Ascensia. The ATLANTIC DIP program was funded through an initial program grant and several subsequent project grants supporting PPC and GDM screening evaluation from the HRB.

The views expressed in the trials are those of the authors and coauthors and were not influenced by the HRB, Merck, or Ascensia.

Duality of Interest. F.D. has received speaker’s honoraria from Novo Nordisk, Sanofi, Menerini, Merck, and Dexcom. No other potential conflicts of interest relevant to this article were reported.

Handling Editors. The journal editors responsible for overseeing the review of the manuscript were Steven E. Kahn and Camille E. Powe.