Dolutegravir use at conception: Additional surveillance data from Botswana Mmakgomo Mimi Raesima, MD, MPH, MHA Botswana Ministry of Health and Wellness On behalf of the Botswana NTD Surveillance Team good morning/afternoon.

Disclosure: No conflicts of interest to declare For disclosure, I have no conflicts of interest to declare.

Botswana poised to provide data on dolutegravir (DTG) Implemented DTG-based ART as first-line therapy in May 2016 Tsepamo study findings raised concerns about a possible association between neural tube defects (NTDs) and DTG use at conception More data needed Botswana was uniquely positioned as it had sufficient numbers of women with DTG exposure at conception In 2018, Botswana was poised to provide data on dolutegravir. The country implemented DTG-based ART as first-line therapy for all adults, including pregnant women, in May 2016. Tsepamo study findings, raised concerns about a possible neural tube defect (NTD) association. More data was needed to understand the possible association. Botswana was uniquely positioned to provide needed data on the safety of DTG as it had sufficient numbers of women with DTG exposure at conception. ______________________________________________________ More than 95% of Botswana’s women deliver in facilities. *Zash, et al. N Engl J Med. 2018;379(10):979-981.

Expanding birth outcomes surveillance Botswana’s Ministry of Health and Wellness (MoHW) initiated the NTD Surveillance Study Conducted in collaboration with: US Centers for Disease Control and Prevention (CDC) – Atlanta and Botswana offices Botswana University of Maryland School of Health Initiative (BUMMHI) Funding support from the US President’s Emergency Plan for AIDS Relief (PEPFAR) Informed by Tsepamo study methods To expand birth surveillance in Botswana, the Ministry of Health and Wellness (MoHW) initiated the NTD Surveillance Study. This study was conducted in collaboration with the US Centers for Disease Control and Prevention (CDC) –Atlanta and Botswana offices, and Botswana University of Maryland School of Health Initiative (BUMMHI). Study funding support came from the US President’s Emergency Plan for AIDS Relief (PEPFAR) and the study’s design was informed by the Tsepamo study methods.

Coverage of birth outcomes surveillance Tsepamo original (8 sites: 45%) After May 2018, coverage of birth outcomes in Botswana increased. Tsepamo added to its original 8 sites.

Coverage of birth outcomes surveillance 72 % Tsepamo original (8 sites: 45%) Tsepamo expansion (10 sites: 27%) = + Tsepamo added another 10 sites; together the original and expansion sites covered 72% of Botswana’s births.

Coverage of birth outcomes surveillance 72 % Tsepamo original (8 sites: 45%) Tsepamo expansion (10 sites: 27%) = + NTD Surveillance Study (22 sites) 19 % = The MoHW NTD Surveillance Study, which was conducted at 22 sites health facilities with highest delivery volume outside of Tsepamo study sites, added coverage for 19% of Botswana’s births.

Coverage of birth outcomes surveillance 72 % Tsepamo original (8 sites: 45%) Tsepamo expansion (10 sites: 27%) = + NTD Surveillance Study (22 sites) 19 % = Together, Tsepamo and MoHW studies provided birth outcome surveillance for more than 90% of births in Botswana. 91 % TOTAL:

Study methods Prospective NTD surveillance Enrolled mothers who delivered infants (live births and stillbirths) >24 weeks gestation at the study sites from October 2018–March 2019 IRB-approved protocol The study used prospective NTD surveillance. Site locations are shown in red on the country map. Prospective surveillance watches for outcomes among a cohort of subjects. In our study, we enrolled mothers who delivered infants (live births and stillbirths) greater than 24 weeks gestation at the study sites from October 2018 through March 2019. The study used an IRB-approved protocol. _________________________________________ Site selection cut-off: deliveries ≥ 100 births/year (using 2017 data) There are 2 back-up slides explaining why retrospective approach not used.

Methods: Prospective Surveillance for All Deliveries At Maternity Unit Examination of infant Completion of Coversheet and Maternal Obstetrical Record + MIDWIFE Delivery The next several slides show the surveillance procedures. At the maternity unit, after delivery, midwives systematically examined the infant. They then completed study coversheets that asked for specific HIV information and documented the head, neck and spine examination, as this was not routinely documented in the standard ANC/delivery Obstetrical Record.

Methods: Prospective Surveillance for All Deliveries At Maternity Unit + MIDWIFE Delivery Examination of infant Completion of Coversheet and Maternal Obstetrical Record Abstraction of data and entry into electronic data collection tool Designated facility Study Surveillance Assistants abstracted data from the coversheet and Obstetrical Record and entered it into the REDCap electronic data base via a tablet. SURVEILLANCE ASSISTANT Review of de-identified information and classification of the suspected NTD

Methods: Prospective Surveillance for Suspected NTDs At Maternity Unit + MIDWIFE Suspected NTD Delivery Examination of infant Completion of Coversheet and Maternal Obstetrical Record Consent for photographs and abstraction of additional information Abstraction of data and entry into electronic data collection tool If an infant with a suspected NTD was delivered, in addition to standard documentation done for all deliveries, maternal consent was requested to photograph the infant and extra information was abstracted from the mother’s Obstetrical Record. SURVEILLANCE ASSISTANT

Methods: Prospective Surveillance for Suspected NTDs At Maternity Unit + MIDWIFE Suspected NTD Delivery Examination of infant Completion of Coversheet and Maternal Obstetrical Record Consent for photographs and abstraction of additional information Abstraction of data and entry into electronic data collection tool Suspected NTDs were reviewed and classified by the clinical geneticist at CDC-Atlanta, who was blinded to ART exposure and other clinical history. SURVEILLANCE ASSISTANT Review of de-identified information and classification of the suspected NTD Review of de-identified information and classification of the suspected NTD

Methods: Prospective Surveillance for All Deliveries At Maternity Unit + MIDWIFE Delivery Examination of infant Completion of Coversheet and Maternal Obstetrical Record Abstraction of data and entry into electronic data collection tool Teams of the study staff regularly visited sites to thoroughly review documents and assure data quality. SURVEILLANCE ASSISTANT Review of de-identified information and classification of the suspected NTD Data quality review for deliveries at all sites

Methods: Exposure definition Exposure of interest = ART at conception HIV-positive women categorized as ART: exposed/not exposed/unknown Last menstrual period used to determine periconceptional time frame Non-DTG based regimen EFV-based regimen Non- EFV-based regimen DTG-based regimen HIV-positive women exposed to ART at conception The study exposure of interest was ART at conception. HIV-positive women were categorized as ART exposed, not exposed, or unknown . The ART regimen categories were DTG-based, non-DTG-based and efavirenz-based, shown on the right. The mother’s last menstrual period was used to determine the periconception time frame.

Methods: Outcome definition Outcome of interest = NTD classification Not NTD — photo/description/drawing sufficient to rule out NTD Possible — description/drawing insufficient to determine NTD status Probable — clear description/drawing and indication of surgical referral, if live-born Confirmed — photo or clear description/drawing and/or surgical records, if live-born The outcome of interest was NTD classification. The four NTD categories were: Not an NTD. Possible NTD--applied when the description or drawing was insufficient to determine NTD status. Probable and confirmed NTD classifications were made when detailed documentation of the defect was provided; but a confirmed NTD had photos or surgical records available.

Results: Pregnancies by HIV and ART status in 22 surveillance sites Between October 2018 and March 2019, our surveillance system captured 3076 pregnancies who delivered at the 22 sites.

Results: Pregnancies by HIV and ART status in 22 surveillance sites 742 (24%) with HIV-positive mother 2328 (76%) with HIV-negative mother 6 (<0.1%) with mother’s HIV status unknown 742 mothers (24%) were HIV positive. 2328 mothers were HIV negative and 6 mothers had unknown HIV status.

Results: Pregnancies by HIV and ART status in 22 surveillance sites 742 (24%) with HIV-positive mother 2328 (76%) with HIV-negative mother 6 (<0.1%) with mother’s HIV status unknown 544 (73%) with HIV-positive mother on ART at conception 176 (24%) with HIV- positive mother not on ART at conception 22 (3%) with HIV-positive mother with unknown ART status at conception Of the 742 HIV positive mothers, 544, about three-quarters, were on ART at conception.

Results: Pregnancies by HIV and ART status in 22 surveillance sites 742 (24%) with HIV-positive mother 2328 (76%) with HIV-negative mother 6 (<0.1%) with mother’s HIV status unknown 544 (73%) with HIV-positive mother on ART at conception 176 (24%) with HIV- positive mother not on ART at conception 22 (3%) with HIV-positive mother with unknown ART status at conception Of the 544 HIV-positive mothers who were on ART at conception, 152, or about 28% were on a DTG-based ART regimen at conception. 381 mothers were on non-DTG regimens and 11 were on unknown regimens at conception. 152 (28%) on DTG-based regimen at conception 381 (70%) on non-DTG- based regimen at conception 11 (2%) on unknown regimen at conception

Results: Pregnancies by HIV and ART status in 22 surveillance sites 742 (24%) with HIV-positive mother 2328 (76%) with HIV-negative mother 6 (<0.1%) with mother’s HIV status unknown 544 (73%) with HIV-positive mother on ART at conception 176 (24%) with HIV- positive mother not on ART at conception 22 (3%) with HIV-positive mother with unknown ART status at conception Of those on non-DTG based regimens at conception, 261 were on efavirenz-based regimens. 152 (28%) on DTG-based regimen at conception 381 (70%) on non-DTG- based regimen at conception 11 (2%) on unknown regimen at conception 261 (69%) on EFV-based regimen at conception 120 (31%) on non-EFV-based regimen at conception

Results: Pregnancies by HIV and ART status in 22 surveillance sites 742 (24%) with HIV-positive mother 2328 (76%) with HIV-negative mother 6 (<0.1%) with mother’s HIV status unknown 544 (73%) with HIV-positive mother on ART at conception 176 (24%) with HIV- positive mother not on ART at conception 22 (3%) with HIV-positive mother with unknown ART status at conception The colored boxes on this slide represent the different exposure groups that we analyzed. 152 (28%) on DTG-based regimen at conception 381 (70%) on non-DTG- based regimen at conception 11 (2%) on unknown regimen at conception 261 (69%) on EFV-based regimen at conception 120 (31%) on non-EFV-based regimen at conception

Results: Six suspected NTDs identified The surveillance system identified 6 suspected NTDs within the participating facilities. These cases were reviewed by the CDC clinical geneticist. Determinations were:

Results: Six suspected NTDs identified 2 not NTDs: sacral dimples 2 were not NTDs, both cases were sacral dimples.

Results: Six suspected NTDs identified 2 not NTDs: sacral dimples 1 possible NTD: unspecified lesion on side of head 1 possible NTD, which was an unspecified lesion on the side of the head.

Results: Six suspected NTDs identified 2 not NTDs: sacral dimples 1 possible NTD: unspecified lesion on side of head 2 probable NTDs: 1 spina bifida and 1 spina bifida with frontal encephalocele 2 probable NTDs, one spina bifida alone; and one spina bifida with frontal encephalocele.

Results: Six suspected NTDs identified 2 not NTDs: sacral dimples 1 possible NTD: unspecified lesion on side of head 2 probable NTDs: 1 spina bifida and 1 spina bifida with frontal encephalocele 1 confirmed NTD: spina bifida And 1 confirmed NTD, spina bifida.

Results: Six suspected NTDs identified 2 not NTDs: sacral dimples 1 possible NTD: unspecified lesion on side of head 2 probable NTDs: 1 spina bifida and 1 spina bifida with frontal encephalocele 1 confirmed NTD: spina bifida Probable and confirmed NTD cases (n=3) were included in the analysis The two probable and one confirmed NTD cases were the outcomes included in the analysis.

Results: NTD prevalence by maternal exposure Percentage (95% CI) with NTD Number of NTDs 1 2 Number of Exposures 152 381 261 2328 % with NTD (95% CI) 0.66% (0.02%, 3.69%) 0 (0, 0.79%) 0 (0, 1.15%) 0.09% (0.01%, 0.31%) Prevalence Difference (95% CI) ref 0.66% (-0.73%, 4.16%) 0.66% (-1.25%, 4.16%) 0.58% (-0.10%, 4.10%) Of the three (3) NTDs included in the final analysis. One occurred among the 152 DTG-exposed HIV-positive mothers. The other two NTDs occurred among the 2328 HIV-negative mothers. Slide adapted from Rebecca Zash

Results: NTD prevalence by maternal exposure DTG 0.66% (1/152) HIV-negative 0.09% (2/2328) Percentage (95% CI) with NTD Number of NTDs 1 2 Number of Exposures 152 381 261 2328 % with NTD (95% CI) 0.66% (0.02%, 3.69%) 0 (0, 0.79%) 0 (0, 1.15%) 0.09% (0.01%, 0.31%) Prevalence Difference (95% CI) ref 0.66% (-0.73%, 4.16%) 0.66% (-1.25%, 4.16%) 0.58% (-0.10%, 4.10%) The NTD prevalence among the DTG-exposed HIV-positive mothers was 0.66% (with 95% Confidence Interval: 0.02% – 3.69%). NTD prevalence among the HIV-negative mothers was 0.09% (with 95% Confidence Interval: 0.01% – 0.31%). Slide adapted from Rebecca Zash

Consideration: How do we monitor birth outcomes? As the dolutegravir-based ART regimen (TLD) is rolled out, countries will be interested in how best to monitor birth outcomes.

Consideration: How do we monitor birth outcomes? Surveillance Pharmacovigilance They will consider two approaches, surveillance and pharmacovigilance. To choose well, decision makers will need to understand and consider the differences between these two approaches.

Consideration: Surveillance Surveillance is a reliable method to establish the DTG-NTD association Data collection methodology is continuous and systematic Provides a denominator to estimate the magnitude and scope of birth defects Starting point = outcome (delivery, NTD) Requires large sample size to detect rare outcome of birth defects Complex and resource-intensive Surveillance is a reliable method to establish the DTG-NTD association. Its advantages include data collection methodology that is continuous and systematic; the approach provides a denominator to estimate the magnitude and scope of birth defects, and the starting point is the outcome (delivery or NTD). Surveillance can be challenging because it requires a large sample size to detect rare outcomes of birth defects; it is complex and resource-intensive.* ____________________________________________ *Back-up slide “Surveillance is complex/resource-intensive” lists components of surveillance.

Consideration: Pharmacovigilance Pharmacovigilance is a common post-marketing approach to capture adverse drug effects No comparison group Not designed to collect data from maternity wards Starting point = exposure (DTG at conception) Pharmacovigilance is a common post-marketing approach to capture adverse drug effects. Pharmacovigilance provides no comparison group, meaning that no information would be collected on HIV-negative women. It is not designed to collect data from maternity wards because its starting point is the exposure (DTG at conception), not the NTD outcome.

Conclusions Small increased NTD prevalence among infants born to HIV-positive mothers on DTG at conception compared to infants born to HIV-negative mothers Study was conducted for only 6 months, resulting in small number of deliveries Data suggest that NTD risk with DTG exposure at conception remains <1%, consistent with Tsepamo study findings reported in 2018 In conclusion, we found a small increased NTD prevalence among infants born to HIV-positive mothers on DTG at conception compared to infants born to HIV-negative mothers. The study was conducted for only 6 months, resulting in a small number of deliveries. Consequently, our NTD prevalence estimates have wide confidence intervals. However, our data suggest that NTD risk with DTG exposure at conception remains <1%, consistent with the Tsepamo study findings reported in 2018.

Acknowledgements Botswana MoHW: CDC Atlanta: CDC Botswana BUMMHI Eldah Dintwa Chipo Petlo Sifelani Malima Omphemetse Mmunyane Kelame Kefitlhile Thusoetsile Modise CDC Botswana Vasavi Thomas Michelle Roland Catherine Motswere-Chirwa Kunle Dare Shifawu Odunsi Mpho Letebele CDC Atlanta: Sara Forhan Dhelia Williamson Elizabeth Rabold Sarah Tinker Cynthia Moore Surbhi Modi BUMMHI Gadzikanani Gokatweng Chibuike Mark Ogbuabo Ndwapi Ndwapi I would like to acknowledge my colleagues from MoHW, CDC Botswana and Atlanta, and BUMMHI for making this study possible.

Ke a leboga! Thank you!

EXTRA SLIDES

Retrospective study not feasible Prospective 2016 2017 2018 Using standard obstetrical records: Unreliable outcome data unable to validate NTDs Missing records, unreliable exams, missing exams, unable to validate Incomplete exposure data—ART regimen at conception not captured In the study planning stages, a review of the standard obstetrical records from more than 200 deliveries at 4 facilities showed that a retrospective study would not be feasible. We found unreliable outcome data, thus we would be unable to validate NTDs; exposure data--specific ART regimen and date of initiation--was incomplete.

Prospective approach chosen Retrospective Prospective 2016 2017 2018 Reasons: Able to capture needed data elements with additional tool Provides an accurate denominator Can compare NTD ratios among other groups of women (e.g., HIV negative and HIV-positive on other ARVs) Thus a prospective approach was chosen. The needed data elements were captured with an additional data collection tool. Moreover, the prospective approach provided an accurate denominator and the ability to compare NTD ratios among other groups of women, for example between HIV-negative and HIV-positive mothers on ARVs other than DTG.

Would the results be different if the analysis had included the possible NTD case? Possible NTD case occurred in an HIV-negative woman We conducted a sensitivity analysis to assess how inclusion of the possible NTD would affect the prevalence difference estimate and found there was no meaningful change The possible NTD case occurred in an HIV-negative woman. We conducted a sensitivity analysis to assess how inclusion of the possible NTD would affect the prevalence difference estimate and found there was no meaningful change.

Birth surveillance is complex and resource-intensive Designated leadership New data tools Midwife training Cadre dedicated to data collection Commodities Logistics Ongoing oversight for procedural and data quality Data collection system We learned that setting up a birth surveillance system, even at a few sites, is complex and resource-intensive. It required significant planning, coordination, and monetary and human resources. Among the needed components for such a system are: Designated leadership New data tools to capture exposure and outcome-related information may be needed Midwife training A cadre dedicated to data collection Commodities (such as storage files, collection forms, and tablets) Logistics Ongoing oversight for procedural and data quality and a Data collection system