Four statistical frameworks for assessing an immune correlate of protection (surrogate endpoint) from a randomized, controlled, vaccine efficacy trial

A central goal of vaccine research is to characterize and validate immune correlates of protection (CoPs). In addition to helping elucidate immunological mechanisms, a CoP can serve as a valid surrogate endpoint for an infectious disease clinical outcome and thus qualifies as a primary endpoint for vaccine authorization or approval without requiring resource-intensive randomized, controlled phase 3 trials. Yet, it is challenging to persuasively validate a CoP, because a prognostic immune marker can fail as a reliable basis for predicting/inferring the level of vaccine efficacy against a clinical outcome, and because the statistical analysis of phase 3 trials only has limited capacity to disentangle association from cause. Moreover, the multitude of statistical methods garnered for CoP evaluation in phase 3 trials renders the comparison, interpretation, and synthesis of CoP results challenging. Toward promoting broader harmonization and standardization of CoP evaluation, this article summarizes four complementary statistical frameworks for evaluating CoPs in a phase 3 trial, focusing on the frameworks' distinct scientific objectives as measured and communicated by distinct causal vaccine efficacy parameters. Advantages and disadvantages of the frameworks are considered, dependent on phase 3 trial context, and perspectives are offered on how the frameworks can be applied and their results synthesized.

Durability of Protection Against COVID-19 Through the Delta Surge for the NVX-CoV2373 Vaccine

BACKGROUND

Protein-based vaccines for COVID-19 provide a traditional vaccine platform with long-lasting protection for non-SARS-CoV-2 pathogens and may complement messenger RNA vaccines as a booster dose. While NVX-CoV2373 showed substantial early efficacy, the durability of protection has not been delineated.

METHODS

The PREVENT-19 vaccine trial employed a blinded crossover design; the original placebo arm received NVX-CoV2373 after efficacy was established. Using novel statistical methods that integrate surveillance data of circulating strains with post-crossover cases, we estimated placebo-controlled vaccine efficacy and durability of NVX-CoV2373 against both pre-Delta and Delta strains of SARS-CoV-2.

RESULTS

Vaccine efficacy against pre-Delta strains of COVID-19 was 89% (95% CI: 75%, 95%) and 87% (72%, 94%) at 0 and 90 days after 2 doses of NVX-CoV2373, respectively, with no evidence of waning (p=0.93). Vaccine efficacy against the Delta strain was 88% (71%, 95%), 82% (56%, 92%), and 77% (44%, 90%) at 40, 120, and 180 days, respectively, with evidence of waning (p<0.01). In sensitivity analyses, the estimated Delta vaccine efficacy at 120 days ranged from 66% (15%, 86%) to 89% (74%, 95%) per various assumptions of the surveillance data.

CONCLUSION

NVX-CoV2373 has high initial efficacy against pre-Delta and Delta strains of COVID-19 with little evidence of waning for pre-Delta strains through 90 days and moderate waning against Delta strains over 180 days.

A controlled effects approach to assessing immune correlates of protection

An immune correlate of risk (CoR) is an immunologic biomarker in vaccine recipients associated with an infectious disease clinical endpoint. An immune correlate of protection (CoP) is a CoR that can be used to reliably predict vaccine efficacy (VE) against the clinical endpoint and hence is accepted as a surrogate endpoint that can be used for accelerated approval or guide use of vaccines. In randomized, placebo-controlled trials, CoR analysis is limited by not assessing a causal vaccine effect. To address this limitation, we construct the controlled risk curve of a biomarker, which provides the causal risk of an endpoint if all participants are assigned vaccine and the biomarker is set to different levels. Furthermore, we propose a causal CoP analysis based on controlled effects, where for the important special case that the biomarker is constant in the placebo arm, we study the controlled vaccine efficacy curve that contrasts the controlled risk curve with placebo arm risk. We provide identification conditions and formulae that account for right censoring of the clinical endpoint and two-phase sampling of the biomarker, and consider G-computation estimation and inference under a semiparametric model such as the Cox model. We add modular approaches to sensitivity analysis that quantify robustness of CoP evidence to unmeasured confounding. We provide an application to two phase 3 trials of a dengue vaccine indicating that controlled risk of dengue strongly varies with 50$\%$ neutralizing antibody titer. Our work introduces controlled effects causal mediation analysis to immune CoP evaluation.

Stochastic Interventional Vaccine Efficacy and Principal Surrogate Analyses of Antibody Markers as Correlates of Protection against Symptomatic COVID-19 in the COVE mRNA-1273 Trial

The COVE trial randomized participants to receive two doses of mRNA-1273 vaccine or placebo on Days 1 and 29 (D1, D29). Anti-SARS-CoV-2 Spike IgG binding antibodies (bAbs), anti-receptor binding domain IgG bAbs, 50% inhibitory dilution neutralizing antibody (nAb) titers, and 80% inhibitory dilution nAb titers were measured at D29 and D57. We assessed these markers as correlates of protection (CoPs) against COVID-19 using stochastic interventional vaccine efficacy (SVE) analysis and principal surrogate (PS) analysis, frameworks not used in our previous COVE immune correlates analyses. By SVE analysis, hypothetical shifts of the D57 Spike IgG distribution from a geometric mean concentration (GMC) of 2737 binding antibody units (BAU)/mL (estimated vaccine efficacy (VE): 92.9% (95% CI: 91.7%, 93.9%)) to 274 BAU/mL or to 27,368 BAU/mL resulted in an overall estimated VE of 84.2% (79.0%, 88.1%) and 97.6% (97.4%, 97.7%), respectively. By binary marker PS analysis of Low and High subgroups (cut-point: 2094 BAU/mL), the ignorance interval (IGI) and estimated uncertainty interval (EUI) for VE were [85%, 90%] and (78%, 93%) for Low compared to [95%, 96%] and (92%, 97%) for High. By continuous marker PS analysis, the IGI and 95% EUI for VE at the 2.5th percentile (519.4 BAU/mL) vs. at the 97.5th percentile (9262.9 BAU/mL) of D57 Spike IgG concentration were [92.6%, 93.4%] and (89.2%, 95.7%) vs. [94.3%, 94.6%] and (89.7%, 97.0%). Results were similar for other D29 and D57 markers. Thus, the SVE and PS analyses additionally support all four markers at both time points as CoPs.

Optimizing Deep Learning-Based Segmentation of Densely Packed Cells using Cell Surface Markers

Background

Spatial molecular profiling depends on accurate cell segmentation. Identification and quantitation of individual cells in dense tissues, e.g. highly inflamed tissue caused by viral infection or immune reaction, remains a challenge.

Methods

We first assess the performance of 18 deep learning-based cell segmentation models, either pre-trained or trained by us using two public image sets, on a set of immunofluorescence images stained with immune cell surface markers in skin tissue obtained during human herpes simplex virus (HSV) infection. We then further train eight of these models using up to 10,000+ training instances from the current image set. Finally, we seek to improve performance by tuning parameters of the most successful method from the previous step.

Results

The best model before fine-tuning achieves a mean Average Precision (mAP) of 0.516. Prediction performance improves substantially after training. The best model is the cyto model from Cellpose. After training, it achieves an mAP of 0.694; with further parameter tuning, the mAP reaches 0.711.

Conclusion

Selecting the best model among the existing approaches and further training the model with images of interest produce the most gain in prediction performance. The performance of the resulting model compares favorably to human performance. The imperfection of the final model performance can be attributed to the moderate signal-to-noise ratio i the imageset.

Examining protective effects of SARS-CoV-2 neutralizing antibodies after vaccination or monoclonal antibody administration

While new vaccines for SARS-CoV-2 are authorized based on neutralizing antibody (nAb) titer against emerging variants of concern, an analogous pathway does not exist for preventative monoclonal antibodies. In this work, nAb titers were assessed as correlates of protection against COVID-19 in the casirivimab + imdevimab monoclonal antibody (mAb) prevention trial (ClinicalTrials.gov #NCT4452318) and in the mRNA-1273 vaccine trial (ClinicalTrials.gov #NCT04470427). In the mAb trial, protective efficacy of 92% (95% confidence interval (CI): 84%, 98%) is associated with a nAb titer of 1000 IU50/ml, with lower efficacy at lower nAb titers. In the vaccine trial, protective efficacies of 93% [95% CI: 91%, 95%] and 97% (95% CI: 95%, 98%) are associated with nAb titers of 100 and 1000 IU50/ml, respectively. These data quantitate a nAb titer correlate of protection for mAbs benchmarked alongside vaccine induced nAb titers and support nAb titer as a surrogate endpoint for authorizing new mAbs.

Comparing antibody assays as correlates of protection against COVID-19 in the COVE mRNA-1273 vaccine efficacy trial

The best assay or marker to define mRNA-1273 vaccine-induced antibodies as a correlate of protection (CoP) is unclear. In the COVE trial, participants received two doses of the mRNA-1273 COVID-19 vaccine or placebo. We previously assessed IgG binding antibodies to the spike protein (spike IgG) or receptor binding domain (RBD IgG) and pseudovirus neutralizing antibody 50 or 80% inhibitory dilution titer measured on day 29 or day 57, as correlates of risk (CoRs) and CoPs against symptomatic COVID-19 over 4 months after dose. Here, we assessed a new marker, live virus 50% microneutralization titer (LV-MN50), and compared and combined markers in multivariable analyses. LV-MN50 was an inverse CoR, with a hazard ratio of 0.39 (95% confidence interval, 0.19 to 0.83) at day 29 and 0.51 (95% confidence interval, 0.25 to 1.04) at day 57 per 10-fold increase. In multivariable analyses, pseudovirus neutralization titers and anti-spike binding antibodies performed best as CoRs; combining antibody markers did not improve correlates. Pseudovirus neutralization titer was the strongest independent correlate in a multivariable model. Overall, these results supported pseudovirus neutralizing and binding antibody assays as CoRs and CoPs, with the live virus assay as a weaker correlate in this sample set. Day 29 markers performed as well as day 57 markers as CoPs, which could accelerate immunogenicity and immunobridging studies.

Immune correlates analysis of a phase 3 trial of the AZD1222 (ChAdOx1 nCoV-19) vaccine

In the phase 3 trial of the AZD1222 (ChAdOx1 nCoV-19) vaccine conducted in the U.S., Chile, and Peru, anti-spike binding IgG concentration (spike IgG) and pseudovirus 50% neutralizing antibody titer (nAb ID50) measured four weeks after two doses were assessed as correlates of risk and protection against PCR-confirmed symptomatic SARS-CoV-2 infection (COVID-19). These analyses of SARS-CoV-2 negative participants were based on case-cohort sampling of vaccine recipients (33 COVID-19 cases by 4 months post dose two, 463 non-cases). The adjusted hazard ratio of COVID-19 was 0.32 (95% CI: 0.14, 0.76) per 10-fold increase in spike IgG concentration and 0.28 (0.10, 0.77) per 10-fold increase in nAb ID50 titer. At nAb ID50 below the limit of detection (< 2.612 IU50/ml), 10, 100, and 270 IU50/ml, vaccine efficacy was -5.8% (-651%, 75.6%), 64.9% (56.4%, 86.9%), 90.0% (55.8%, 97.6%) and 94.2% (69.4%, 99.1%). These findings provide further evidence towards defining an immune marker correlate of protection to help guide regulatory/approval decisions for COVID-19 vaccines.

Correlates of Protection, Thresholds of Protection, and Immunobridging among Persons with SARS-CoV-2 Infection

Several studies have shown that neutralizing antibody levels correlate with immune protection from COVID-19 and have estimated the relationship between neutralizing antibodies and protection. However, results of these studies vary in terms of estimates of the level of neutralizing antibodies required for protection. By normalizing antibody titers, we found that study results converge on a consistent relationship between antibody levels and protection from COVID-19. This finding can be useful for planning future vaccine use, determining population immunity, and reducing the global effects of the COVID-19 pandemic.

Immune correlates analysis of the PREVENT-19 COVID-19 vaccine efficacy clinical trial

In the PREVENT-19 phase 3 trial of the NVX-CoV2373 vaccine (NCT04611802), anti-spike binding IgG concentration (spike IgG), anti-RBD binding IgG concentration (RBD IgG), and pseudovirus 50% neutralizing antibody titer (nAb ID50) measured two weeks post-dose two are assessed as correlates of risk and as correlates of protection against COVID-19. Analyses are conducted in the U.S. cohort of baseline SARS-CoV-2 negative per-protocol participants using a case-cohort design that measures the markers from all 12 vaccine recipient breakthrough COVID-19 cases starting 7 days post antibody measurement and from 639 vaccine recipient non-cases. All markers are inversely associated with COVID-19 risk and directly associated with vaccine efficacy. In vaccine recipients with nAb ID50 titers of 50, 100, and 7230 international units (IU50)/ml, vaccine efficacy estimates are 75.7% (49.8%, 93.2%), 81.7% (66.3%, 93.2%), and 96.8% (88.3%, 99.3%). The results support potential cross-vaccine platform applications of these markers for guiding decisions about vaccine approval and use.

Antibody Correlates of Protection From Severe Respiratory Syncytial Virus Disease in a Vaccine Efficacy Trial

Background

Respiratory syncytial virus (RSV) can cause serious lung infections in young children and there is currently no available vaccine.

Methods

We used complementary statistical frameworks to analyze 4 RSV serology measurements in mothers and their infants in South Africa who participated in a phase 3 maternal immunization trial of an RSV F protein nanoparticle vaccine as correlates of risk and of protection against different RSV disease endpoints.

Results

We found evidence to support each antibody measurement-encompassing RSV-neutralizing antibodies and F surface glycoprotein-binding antibodies-as an inverse correlate of risk of RSV-associated acute lower respiratory tract infection with severe hypoxia in at least 1 framework, with vaccine-induced fold-rise from the maternal enrollment to day 14 samples of anti-F immunoglobulin G (IgG) binding antibodies having the most consistent evidence. This evidence includes a significant association of fold-rise anti-F IgG with vaccine efficacy (VE); achieving a baseline covariate-adjusted VE of 75% requires a vaccine-induced maternal anti-F IgG fold-rise of around 16. Neither multivariable logistic regression nor superlearning analyses showed benefit to including multiple time points or assays in the same model, suggesting a parsimonious correlate. Post hoc exploratory analyses supported adherence of vaccine-induced maternal anti-F IgG fold-rise to the Prentice criteria for a valid surrogate endpoint.

Conclusions

Our results suggest that the vaccine induced protective anti-F antibody responses. If this finding is confirmed, VE could potentially be augmented by increasing these responses.

Respiratory syncytial virus prevention within reach: the vaccine and monoclonal antibody landscape

Respiratory syncytial virus is the second most common cause of infant mortality and a major cause of morbidity and mortality in older adults (aged >60 years). Efforts to develop a respiratory syncytial virus vaccine or immunoprophylaxis remain highly active. 33 respiratory syncytial virus prevention candidates are in clinical development using six different approaches: recombinant vector, subunit, particle-based, live attenuated, chimeric, and nucleic acid vaccines; and monoclonal antibodies. Nine candidates are in phase 3 clinical trials. Understanding the epitopes targeted by highly neutralising antibodies has resulted in a shift from empirical to rational and structure-based vaccine and monoclonal antibody design. An extended half-life monoclonal antibody for all infants is likely to be within 1 year of regulatory approval (from August, 2022) for high-income countries. Live-attenuated vaccines are in development for older infants (aged >6 months). Subunit vaccines are in late-stage trials for pregnant women to protect infants, whereas vector, subunit, and nucleic acid approaches are being developed for older adults. Urgent next steps include ensuring access and affordability of a respiratory syncytial virus vaccine globally. This review gives an overview of respiratory syncytial virus vaccines and monoclonal antibodies in clinical development highlighting different target populations, antigens, and trial results.

Immune correlates analysis of the ENSEMBLE single Ad26.COV2.S dose vaccine efficacy clinical trial

Measuring immune correlates of disease acquisition and protection in the context of a clinical trial is a prerequisite for improved vaccine design. We analysed binding and neutralizing antibody measurements 4 weeks post vaccination as correlates of risk of moderate to severe-critical COVID-19 through 83 d post vaccination in the phase 3, double-blind placebo-controlled phase of ENSEMBLE, an international randomized efficacy trial of a single dose of Ad26.COV2.S. We also evaluated correlates of protection in the trial cohort. Of the three antibody immune markers we measured, we found most support for 50% inhibitory dilution (ID50) neutralizing antibody titre as a correlate of risk and of protection. The outcome hazard ratio was 0.49 (95% confidence interval 0.29, 0.81; P = 0.006) per 10-fold increase in ID50; vaccine efficacy was 60% (43%, 72%) at non-quantifiable ID50 (<2.7 IU50 ml-1) and increased to 89% (78%, 96%) at ID50 = 96.3 IU50 ml-1. Comparison of the vaccine efficacy by ID50 titre curves for ENSEMBLE-US, the COVE trial of the mRNA-1273 vaccine and the COV002-UK trial of the AZD1222 vaccine supported the ID50 titre as a correlate of protection across trials and vaccine types.

Vapour pressure deficit determines critical thresholds for global coffee production under climate change

Our understanding of the impact of climate change on global coffee production is largely based on studies focusing on temperature and precipitation, but other climate indicators could trigger critical threshold changes in productivity. Here, using generalized additive models and threshold regression, we investigate temperature, precipitation, soil moisture and vapour pressure deficit (VPD) effects on global Arabica coffee productivity. We show that VPD during fruit development is a key indicator of global coffee productivity, with yield declining rapidly above 0.82 kPa. The risk of exceeding this threshold rises sharply for most countries we assess, if global warming exceeds 2 °C. At 2.9 °C, countries making up 90% of global supply are more likely than not to exceed the VPD threshold. The inclusion of VPD and the identification of thresholds appear critical for understanding climate change impacts on coffee and for the design of adaptation strategies.

The clinical effect of point-of-care HIV diagnosis in infants: a systematic review and meta-analysis

BACKGROUND

Timely diagnosis and treatment of HIV is crucial in HIV-exposed infants to prevent the high rates of mortality seen during the first 2 years of life if HIV is untreated. However, challenges with sample transportation, testing, and result delivery to caregivers have led to long delays in treatment initiation. We aimed to compare the clinical effect of point-of-care HIV testing versus laboratory-based testing (standard of care) in HIV-exposed infants.

METHODS

We did a systematic review and meta-analysis and searched PubMed, MEDLINE, Cochrane Central Register of Controlled Trials, Embase, Conference Proceedings Citation Index-Science, and WHO Global Index Medicus, from Jan 1, 2014, to Aug 31, 2020. Studies were included if they pertained to the use of point-of-care nucleic acid testing for infant HIV diagnosis, had a laboratory-based nucleic acid test as the comparator or standard of care against the index test (same-day point-of-care testing), evaluated clinical outcomes when point-of-care testing was used, and included HIV-exposed infants aged younger than 2 years. Studies were excluded if they did not use a laboratory-based comparator, a nucleic acid test that had been approved by a stringent regulatory authority, or diagnostic-accuracy or performance evaluations (eg, no clinical outcomes included). Reviews, non-research letters, commentaries, and editorials were also excluded. The risk of bias was evaluated using the ROBINS-I framework. Data were extracted from published reports. Data from all studies were analysed using frequency statistics to describe the overall populations evaluated and their results. Key outcomes were time to result delivery and antiretroviral therapy initiation, and proportion of HIV-positive infants initiated on antiretroviral therapy within 60 days after sample collection.

FINDINGS

164 studies were identified by the search and seven were included in the analysis, comprising 37 377 infants in total across 15 countries, including 25 170 (67%) who had point-of-care HIV testing and 12 207 (33%) who had standard-of-care testing. The certainty of evidence was high. Same-day point-of-care testing led to a significantly shorter time between sample collection and result delivery to caregivers compared with standard-of-care testing (median 0 days [95% CI 0-0] vs 35 days [35-37]). Time from sample collection to antiretroviral therapy initiation in infants found to be HIV-positive was significantly lower with point-of-care testing compared with standard of care (median 0 days [95% CI 0-1] vs 40 days [36-44]). When each study's result was weighted equally, 90·3% (95% CI 76·7-96·5) of HIV-positive infants diagnosed using point-of-care testing had started antiretroviral therapy within 60 days of sample collection, compared with only 51·6% (27·1-75·7) who had standard-of-care testing (odds ratio 8·74 [95% CI 6·6-11·6]; p<0·0001).

INTERPRETATION

Overall, the certainty of the evidence in this analysis was rated as high for the primary outcomes related to result delivery and treatment initiation, with no serious risk of bias, inconsistency, indirectness, or imprecision. In HIV-exposed infants, same-day point-of-care HIV testing was associated with significantly improved time to result delivery, time to antiretroviral therapy initiation, and proportion of HIV-positive infants starting antiretroviral therapy within 60 days compared with standard of care.

FUNDING

The Bill & Melinda Gates Foundation.

A controlled effects approach to assessing immune correlates of protection

An immune correlate of risk (CoR) is an immunologic biomarker in vaccine recipients associated with an infectious disease clinical endpoint. An immune correlate of protection (CoP) is a CoR that can be used to reliably predict vaccine efficacy (VE) against the clinical endpoint and hence is accepted as a surrogate endpoint that can be used for accelerated approval or guide use of vaccines. In randomized, placebo-controlled trials, CoR analysis is limited by not assessing a causal vaccine effect. To address this limitation, we construct the controlled risk curve of a biomarker, which provides the causal risk of an endpoint if all participants are assigned vaccine and the biomarker is set to different levels. Furthermore, we propose a causal CoP analysis based on controlled effects, where for the important special case that the biomarker is constant in the placebo arm, we study the controlled vaccine efficacy curve that contrasts the controlled risk curve with placebo arm risk. We provide identification conditions and formulae that account for right censoring of the clinical endpoint and two-phase sampling of the biomarker, and consider G-computation estimation and inference under a semiparametric model such as the Cox model. We add modular approaches to sensitivity analysis that quantify robustness of CoP evidence to unmeasured confounding. We provide an application to two phase 3 trials of a dengue vaccine indicating that controlled risk of dengue strongly varies with 50$\%$ neutralizing antibody titer. Our work introduces controlled effects causal mediation analysis to immune CoP evaluation.

HIV viral load assays when used with whole blood perform well as a diagnostic assay for infants

Objective

Over the past several years, only approximately 50% of HIV-exposed infants received an early infant diagnosis test within the first two months of life. While high attrition and mortality account for some of the shortcomings in identifying HIV-infected infants early and putting them on life-saving treatment, fragmented and challenging laboratory systems are an added barrier. We sought to determine the accuracy of using HIV viral load assays for infant diagnosis of HIV.

Methods

We enrolled 866 Ugandan infants between March–April 2018 for this study after initial laboratory diagnosis. The median age was seven months, while 33% of infants were less than three months of age. Study testing was done using either the Roche or Abbott molecular technologies at the Central Public Health Laboratory. Dried blood spot samples were prepared according to manufacturer-recommended protocols for both the qualitative and quantitative assays. Viral load test samples for the Roche assay were processed using two different buffers: phosphate-buffered saline (PBS: free virus elution viral load protocol [FVE]) and Sample Pre-Extraction Reagent (SPEX: qualitative buffer). Dried blood spot samples were processed for both assays on the Abbott using the manufacturer’s standard infant diagnosis protocol. All infants received a qualitative test for clinical management and additional paired quantitative tests.

Results

858 infants were included in the analysis, of which 50% were female. Over 75% of mothers received antiretroviral therapy, while approximately 65% of infants received infant prophylaxis. The Roche SPEX and Abbott technologies had high sensitivity (>95%) and specificity (>98%). The Roche FVE had lower sensitivity (85%) and viral load values.

Conclusions

To simplify and streamline laboratory practices, HIV viral load may be used to diagnose HIV infection in infants, particularly using the Roche SPEX and Abbott technologies.

Immune Correlates Analysis of a Single Ad26.COV2.S Dose in the ENSEMBLE COVID-19 Vaccine Efficacy Clinical Trial

Anti-spike IgG binding antibody, anti-receptor binding domain IgG antibody, and pseudovirus neutralizing antibody measurements four weeks post-vaccination were assessed as correlates of risk of moderate to severe-critical COVID-19 outcomes through 83 days post-vaccination and as correlates of protection following a single dose of Ad26.COV2.S COVID-19 vaccine in the placebo-controlled phase of ENSEMBLE, an international, randomized efficacy trial. Each marker had evidence as a correlate of risk and of protection, with strongest evidence for 50% inhibitory dilution (ID50) neutralizing antibody titer. The outcome hazard ratio was 0.49 (95% confidence interval 0.29, 0.81; p=0.006) per 10-fold increase in ID50; vaccine efficacy was 60% (43, 72%) at nonquantifiable ID50 (< 2.7 IU50/ml) and rose to 89% (78, 96%) at ID50 = 96.3 IU50/ml. Comparison of the vaccine efficacy by ID50 titer curves for ENSEMBLE-US, the COVE trial of the mRNA-1273 vaccine, and the COV002-UK trial of the AZD1222 vaccine supported consistency of the ID50 titer correlate of protection across trials and vaccine types.

Diagnostic Accuracy of Dried Plasma Spot Specimens for HIV-1 Viral Load Testing: A Systematic Review and Meta-analysis

BACKGROUND

Dried plasma spot specimens may be a viable alternative to traditional liquid plasma in field settings, but the diagnostic accuracy is not well understood.

METHODS

Standard databases (PubMed and Medline), conferences, and gray literature were searched until January 2019. The quality of evidence was evaluated using the Standards for Reporting Studies of Diagnostic Accuracy and Quality Assessment of Diagnostic Accuracy Studies-2 criteria. We used univariate and bivariate random effects models to determine misclassification, sensitivity, and specificity across multiple thresholds, overall and for each viral load technology, and to account for between-study variation.

RESULTS

We identified 23 studies for inclusion in the systematic review that compared the diagnostic accuracy of dried plasma spots with that of plasma. Primary data from 16 of the 23 studies were shared and included in the meta-analysis, representing 18 countries, totaling 1847 paired dried plasma spot:plasma data points. The mean bias of dried plasma spot specimens compared with that of plasma was 0.28 log10 copies/mL, whereas the difference in median viral load was 2.25 log10 copies/mL. More dried plasma spot values were undetectable compared with plasma values (43.6% vs. 29.8%). Analyzing all technologies together, the sensitivity and specificity of dried plasma spot specimens were >92% across all treatment failure thresholds compared and total misclassification <5.4% across all treatment failure thresholds compared. Some technologies had lower sensitivity or specificity; however, the results were typically consistent across treatment failure thresholds.

DISCUSSION

Overall, dried plasma spot specimens performed relatively well compared with plasma with sensitivity and specificity values greater than 90% and misclassification rates less than 10% across all treatment failure thresholds reviewed.

Change Point Inference in the Presence of Missing Covariates for Principal Surrogate Evaluation in Vaccine Trials

We consider the use of threshold-based regression models for evaluating immune response biomarkers as principal surrogate markers of a vaccine's protective effect. Threshold-based regression models, which allow the relationship between a clinical outcome and a covariate to change dramatically across a threshold value in the covariate, have been studied by various authors under fully observed data. Limited research, however, has examined these models in the presence of missing covariates, such as the counterfactual potential immune responses of a participant in the placebo arm of a standard vaccine trial had s/he been assigned to the vaccine arm instead. Based on a hinge model for a threshold effect of the principal surrogate on vaccine efficacy, we develop a regression methodology that consists of two components: (1) The estimated likelihood method is employed to handle missing potential outcomes, and (2) a penalty is imposed on the estimated likelihood to ensure satisfactory finite sample performance. We develop a method that allows joint estimation of all model parameters as well as a two-step method that separates the estimation of the threshold parameter from the rest of the parameters. Stable iterative algorithms are developed to implement the two methods and the asymptotic properties of the proposed estimators are established. In simulation studies, the proposed estimators are shown to have satisfactory finite sample performance. The proposed methods are applied to analyze a real dataset collected from dengue vaccine efficacy trials to predict how vaccine efficacy varies with an individual's potential immune response if receiving vaccine.

Immune correlates analysis of the mRNA-1273 COVID-19 vaccine efficacy clinical trial

In the coronavirus efficacy (COVE) phase 3 clinical trial, vaccine recipients were assessed for neutralizing and binding antibodies as correlates of risk for COVID-19 disease and as correlates of protection. These immune markers were measured at the time of second vaccination and 4 weeks later, with values reported in standardized World Health Organization international units. All markers were inversely associated with COVID-19 risk and directly associated with vaccine efficacy. Vaccine recipients with postvaccination 50% neutralization titers 10, 100, and 1000 had estimated vaccine efficacies of 78% (95% confidence interval, 54 to 89%), 91% (87 to 94%), and 96% (94 to 98%), respectively. These results help define immune marker correlates of protection and may guide approval decisions for messenger RNA (mRNA) COVID-19 vaccines and other COVID-19 vaccines.

Frequent Development of Broadly Neutralizing Antibodies in Early Life in a Large Cohort of Children With Human Immunodeficiency Virus

BACKGROUND

Recent studies have indicated that broadly neutralizing antibodies (bnAbs) in children may develop earlier after human immunodeficiency virus (HIV) infection compared to adults.

METHODS

We evaluated plasma from 212 antiretroviral therapy-naive children with HIV (1-3 years old). Neutralization breadth and potency was assessed using a panel of 10 viruses and compared to adults with chronic HIV. The magnitude, epitope specificity, and immunoglobulin (Ig)G subclass distribution of Env-specific antibodies were assessed using a binding antibody multiplex assay.

RESULTS

One-year-old children demonstrated neutralization breadth comparable to chronically infected adults, whereas 2- and 3-year-olds exhibited significantly greater neutralization breadth (P = .014). Likewise, binding antibody responses increased with age, with levels in 2- and 3-year-old children comparable to adults. Overall, there was no significant difference in antibody specificities or IgG subclass distribution between the pediatric and adult cohorts. It is interesting to note that the neutralization activity was mapped to a single epitope (CD4 binding site, V2 or V3 glycans) in only 5 of 38 pediatric broadly neutralizing samples, which suggests that most children may develop a polyclonal neutralization response.

CONCLUSIONS

These results contribute to a growing body of evidence suggesting that initiating HIV immunization early in life may present advantages for the development of broadly neutralizing antibody responses.

Forward Stepwise Deep Autoencoder-based Monotone Nonlinear Dimensionality Reduction Methods

Dimensionality reduction is an unsupervised learning task aimed at creating a low-dimensional summary and/or extracting the most salient features of a dataset. Principal components analysis (PCA) is a linear dimensionality reduction method in the sense that each principal component is a linear combination of the input variables. To allow features that are nonlinear functions of the input variables, many nonlinear dimensionality reduction methods have been proposed. In this paper we propose novel nonlinear dimensionality reduction methods based on bottleneck deep autoencoders (Kramer, 1991). Our contributions are two-fold: (1) We introduce a monotonicity constraint into bottleneck deep autoencoders for estimating a single nonlinear component and propose two methods for fitting the model. (2) We propose a new, forward stepwise (FS) deep learning architecture for estimating multiple nonlinear components. The former helps extract interpretable, monotone components when the assumption of monotonicity holds, and the latter helps evaluate reconstruction errors in the original data space for a range of components. We conduct numerical studies to compare different model fitting methods and use two real data examples from the studies of human immune responses to HIV to illustrate the proposed methods.

Improving random forest predictions in small datasets from two-phase sampling designs

BACKGROUND

While random forests are one of the most successful machine learning methods, it is necessary to optimize their performance for use with datasets resulting from a two-phase sampling design with a small number of cases-a common situation in biomedical studies, which often have rare outcomes and covariates whose measurement is resource-intensive.

METHODS

Using an immunologic marker dataset from a phase III HIV vaccine efficacy trial, we seek to optimize random forest prediction performance using combinations of variable screening, class balancing, weighting, and hyperparameter tuning.

RESULTS

Our experiments show that while class balancing helps improve random forest prediction performance when variable screening is not applied, class balancing has a negative impact on performance in the presence of variable screening. The impact of the weighting similarly depends on whether variable screening is applied. Hyperparameter tuning is ineffective in situations with small sample sizes. We further show that random forests under-perform generalized linear models for some subsets of markers, and prediction performance on this dataset can be improved by stacking random forests and generalized linear models trained on different subsets of predictors, and that the extent of improvement depends critically on the dissimilarities between candidate learner predictions.

CONCLUSION

In small datasets from two-phase sampling design, variable screening and inverse sampling probability weighting are important for achieving good prediction performance of random forests. In addition, stacking random forests and simple linear models can offer improvements over random forests.

Immune Correlates Analysis of the mRNA-1273 COVID-19 Vaccine Efficacy Trial

BACKGROUND

In the Coronavirus Efficacy (COVE) trial, estimated mRNA-1273 vaccine efficacy against coronavirus disease-19 (COVID-19) was 94%. SARS-CoV-2 antibody measurements were assessed as correlates of COVID-19 risk and as correlates of protection.

METHODS

Through case-cohort sampling, participants were selected for measurement of four serum antibody markers at Day 1 (first dose), Day 29 (second dose), and Day 57: IgG binding antibodies (bAbs) to Spike, bAbs to Spike receptor-binding domain (RBD), and 50% and 80% inhibitory dilution pseudovirus neutralizing antibody titers calibrated to the WHO International Standard (cID50 and cID80). Participants with no evidence of previous SARS-CoV-2 infection were included. Cox regression assessed in vaccine recipients the association of each Day 29 or 57 serologic marker with COVID-19 through 126 or 100 days of follow-up, respectively, adjusting for risk factors.

RESULTS

Day 57 Spike IgG, RBD IgG, cID50, and cID80 neutralization levels were each inversely correlated with risk of COVID-19: hazard ratios 0.66 (95% CI 0.50, 0.88; p=0.005); 0.57 (0.40, 0.82; p=0.002); 0.42 (0.27, 0.65; p<0.001); 0.35 (0.20, 0.61; p<0.001) per 10-fold increase in marker level, respectively, multiplicity adjusted P-values 0.003-0.010. Results were similar for Day 29 markers (multiplicity adjusted P-values <0.001-0.003). For vaccine recipients with Day 57 reciprocal cID50 neutralization titers that were undetectable (<2.42), 100, or 1000, respectively, cumulative incidence of COVID-19 through 100 days post Day 57 was 0.030 (0.010, 0.093), 0.0056 (0.0039, 0.0080), and 0.0023 (0.0013, 0.0036). For vaccine recipients at these titer levels, respectively, vaccine efficacy was 50.8% (-51.2, 83.0%), 90.7% (86.7, 93.6%), and 96.1% (94.0, 97.8%). Causal mediation analysis estimated that the proportion of vaccine efficacy mediated through Day 29 cID50 titer was 68.5% (58.5, 78.4%).

CONCLUSIONS

Binding and neutralizing antibodies correlated with COVID-19 risk and vaccine efficacy and likely have utility in predicting mRNA-1273 vaccine efficacy against COVID-19.

TRIAL REGISTRATION NUMBER

COVE ClinicalTrials.gov number, NCT04470427.

A Deferred-Vaccination Design to Assess Durability of COVID-19 Vaccine Effect After the Placebo Group Is Vaccinated

Multiple candidate vaccines to prevent COVID-19 have entered large-scale phase 3 placebo-controlled randomized clinical trials, and several have demonstrated substantial short-term efficacy. At some point after demonstration of substantial efficacy, placebo recipients should be offered the efficacious vaccine from their trial, which will occur before longer-term efficacy and safety are known. The absence of a placebo group could compromise assessment of longer-term vaccine effects. However, by continuing follow-up after vaccination of the placebo group, this study shows that placebo-controlled vaccine efficacy can be mathematically derived by assuming that the benefit of vaccination over time has the same profile for the original vaccine recipients and the original placebo recipients after their vaccination. Although this derivation provides less precise estimates than would be obtained by a standard trial where the placebo group remains unvaccinated, this proposed approach allows estimation of longer-term effect, including durability of vaccine efficacy and whether the vaccine eventually becomes harmful for some. Deferred vaccination, if done open-label, may lead to riskier behavior in the unblinded original vaccine group, confounding estimates of long-term vaccine efficacy. Hence, deferred vaccination via blinded crossover, where the vaccine group receives placebo and vice versa, would be the preferred way to assess vaccine durability and potential delayed harm. Deferred vaccination allows placebo recipients timely access to the vaccine when it would no longer be proper to maintain them on placebo, yet still allows important insights about immunologic and clinical effectiveness over time.

Immunobridging efficacy of a tetravalent dengue vaccine against dengue and against hospitalized dengue from children/adolescents to adults in highly endemic countries

BACKGROUND

The recombinant tetravalent live-attenuated dengue vaccine based on the YF 17D vaccine virus backbone (CYD-TDV) demonstrated vaccine efficacy (VE) against symptomatic, virologically confirmed dengue of any serotype from month 13 to month 25 (VCD-DENV-AnyM13→M25) in the CYD14 (2-14-y-olds) and CYD15 (9-16-y-olds) phase 3 trials. Fifty percent plaque reduction neutralization test (PRNT50) titers are a potential surrogate for immunobridging VE to adults.

METHODS

Using PRNT50 calibration datasets, we applied immunobridging approaches using baseline and/or M13 PRNT50 titers to estimate VE against VCD-DENV-AnyM0→M25 and against hospitalized VCD (HVCD)-DENV-AnyM0→M72 in hypothetical 18-45-y-old and 46-50-y-old CYD14 and CYD15 cohorts.

RESULTS

Baseline and M13 geometric mean PRNT50 titers were greater in 18-45-y-olds and in 46-50-y-olds vs 9-16-y-olds for most comparisons. Estimated VE (95% CIs against VCD-DENV-AnyM0→M25 ranged from 75.3% to 90.9% (52.5% to 100%) for 18-45-y-olds and 74.8% to 92.0% (53.4% to 100%) for 46-50-y-olds. Estimated VE (95% CIs) against HVCD-DENV-AnyM0→M72 ranged from 58.8% to 78.1% (40.9 to 98.9%) for 18-45-y-olds and 57.2% to 78.4% (40.5 to 97.6%) for 46-50-y-olds. Corresponding predictions among baseline-seropositive individuals yielded comparable or higher VE estimates.

CONCLUSIONS

VE M0→M25 against DENV-Any and VE against HVCD-DENV-AnyM0→M72 are both expected to be higher in 18-45 and 46-50-y-olds vs CYD14 and CYD15 9-16-y-olds.

Four-parameter paired response curve for serial dilution assays

Newer immunoassay platforms offer improved signal-to-noise ratio but are more expensive. Thus, it is more cost efficient to perform these assays at a few selected, rather than a full series of, sample dilutions. We propose a new four-parameter paired response curve to model the relationship between assay outcomes from two sample dilutions and study likelihood-based inference. Given a fitted paired response curve, we can predict assay outcomes for de novo dilutions of samples, which enables cross-protocol comparison of immune response biomarkers even when different protocols use different sample dilutions. Numerical studies on both simulated and real data are presented.

Fast Grid Search and Bootstrap-based Inference for Continuous Two-phase Polynomial Regression Models

Two-phase polynomial regression models (Robison, 1964; Fuller, 1969; Gallant and Fuller, 1973; Zhan et al., 1996) are widely used in ecology, public health, and other applied fields to model nonlinear relationships. These models are characterized by the presence of threshold parameters, across which the mean functions are allowed to change. That the threshold is a parameter of the model to be estimated from the data is an essential feature of two-phase models. It distinguishes them, and more generally, multi-phase models, from the spline models and has profound implications for both computation and inference for the models. Estimation of two-phase polynomial regression models is a non-convex, non-smooth optimization problem. Grid search provides high quality solutions to the estimation problem, but is very slow when done by brute force. Building upon our previous work on piecewise linear two-phase regression models estimation, we develop fast grid search algorithms for two-phase polynomial regression models and demonstrate their performance. Furthermore, we develop bootstrap-based pointwise and simultaneous confidence bands for mean functions. Monte Carlo studies are conducted to demonstrate the computational and statistical properties of the proposed methods. Three real datasets are used to help illustrate the application of two-phase models, with special attention on model choice.

Threshold-based subgroup testing in logistic regression models in two-phase sampling designs

The effect of treatment on binary disease outcome can differ across subgroups characterized by other covariates. Testing for the existence of subgroups that are associated with heterogeneous treatment effects can provide valuable insight regarding the optimal treatment recommendation in practice. Our research in this paper is motivated by the question of whether host genetics could modify a vaccine's effect on HIV acquisition risk. To answer this question, we used data from an HIV vaccine trial with a two-phase sampling design and developed a general threshold-based model framework to test for the existence of subgroups associated with the heterogeneity in disease risks, allowing for subgroups based on multivariate covariates. We developed a testing procedure based on maximum of likelihood-ratio statistics over change planes and demonstrated its advantage over alternative methods. We further developed the testing procedure to account for bias sampling of expensive (i.e. resource-intensive to measure) covariates through the incorporation of inverse probability weighting techniques. We used the proposed method to analyze the motivating HIV vaccine trial data. Our proposed testing procedure also has broad applications in epidemiological studies for assessing heterogeneity in disease risk with respect to univariate or multivariate predictors.

Clinical Endpoints for Evaluating Efficacy in COVID-19 Vaccine Trials

Several vaccine candidates to protect against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection or coronavirus disease 2019 (COVID-19) have entered or will soon enter large-scale, phase 3, placebo-controlled randomized clinical trials. To facilitate harmonized evaluation and comparison of the efficacy of these vaccines, a general set of clinical endpoints is proposed, along with considerations to guide the selection of the primary endpoints on the basis of clinical and statistical reasoning. The plausibility that vaccine protection against symptomatic COVID-19 could be accompanied by a shift toward more SARS-CoV-2 infections that are asymptomatic is highlighted, as well as the potential implications of such a shift.

Assessing Durability of Vaccine Effect Following Blinded Crossover in COVID-19 Vaccine Efficacy Trials

BACKGROUND

Several candidate vaccines to prevent COVID-19 disease have entered large-scale phase 3 placebo-controlled randomized clinical trials and some have demonstrated substantial short-term efficacy. Efficacious vaccines should, at some point, be offered to placebo participants, which will occur before long-term efficacy and safety are known.

METHODS

Following vaccination of the placebo group, we show that placebo-controlled vaccine efficacy can be derived by assuming the benefit of vaccination over time has the same profile for the original vaccine recipients and the placebo crossovers. This reconstruction allows estimation of both vaccine durability and potential vaccine-associated enhanced disease.

RESULTS

Post-crossover estimates of vaccine efficacy can provide insights about durability, identify waning efficacy, and identify late enhancement of disease, but are less reliable estimates than those obtained by a standard trial where the placebo cohort is maintained. As vaccine efficacy estimates for post-crossover periods depend on prior vaccine efficacy estimates, longer pre-crossover periods with higher case counts provide better estimates of late vaccine efficacy. Further, open-label crossover may lead to riskier behavior in the immediate crossover period for the unblinded vaccine arm, confounding vaccine efficacy estimates for all post-crossover periods.

CONCLUSIONS

We advocate blinded crossover and continued follow-up of trial participants to best assess vaccine durability and potential delayed enhancement of disease. This approach allows placebo recipients timely access to the vaccine when it would no longer be proper to maintain participants on placebo, yet still allows important insights about immunological and clinical effectiveness over time.

Rapid Collection of Human Rectal Secretions Using OriCol Devices Is Suitable for Measurement of Mucosal Ig without Blood Contamination

Measurements of IgG and IgA in human rectal secretions are used to evaluate the Abs elicited by HIV vaccines or the bioaccumulation following immunoprophylaxis at the sites of HIV exposure. To improve sampling methods and tolerability of the procedure, we optimized a balloon device (OriCol) for rectal microbiome sampling requiring 10 second inflation and compared this method to a 5 minute collection using sponges. Lubrication of the device did not interfere with IgG, IgA, or hemoglobin ELISA. Lubricated OriCols inflated to 30 cc minimized hemoglobin contamination (<4.68 ng/ml) compared with collections with two sponge types (Weck-Cel: 267.2 ng/ml, p < 0.0001; and Merocel: 59.38 ng/ml, p = 0.003). Median human serum albumin for OriCols was 14.9 μg/ml, whereas Merocels and Weck-Cels were 28.57 μg/ml (p = 0.0005) and 106.2 μg/ml (p = 0.0002), respectively. Consistent with reduced systemic contamination, the median IgG measured in OriCol-collected rectal secretions (986 ng) was lower than secretions from sponges (Weck-Cel: 8588 ng, p < 0.0001; Merocel: 2509 ng, p = 0.0389). The median IgA yield of samples using the OriCol method (75,253 ng) was comparable to that using Merocel (71,672 ng; p = 0.6942) but significantly higher than Weck-Cel sponges (16,173 ng, p = 0.0336). Median recovery volumes for OriCols were 800 μl, whereas Merocels and Weck-Cels were 615 μl (p = 0.0010) and 655 μl (p = 0.0113), respectively. The balloon device was acceptable among 23 participants, as 85.1% experiencing their first collection ranked it as "seven: acceptable - a lot" or "six: acceptable - somewhat" in a seven-point Likert scale. Therefore, lubricated OriCols inflated to 30 cc allowed for a rapid, well-tolerated, blood-free collection of human rectal secretions.

Microneutralization assay titer correlates analysis in two phase 3 trials of the CYD-TDV tetravalent dengue vaccine in Asia and Latin America

We previously showed that Month 13 50% plaque reduction neutralization test (PRNT50) neutralizing antibody (nAb) titers against dengue virus (DENV) correlated with vaccine efficacy (VE) of CYD-TDV against symptomatic, virologically-confirmed dengue (VCD) in the CYD14 and CYD15 Phase 3 trials. While PRNT is the gold standard nAb assay, it is time-consuming and costly. We developed a next-generation high-throughput microneutralization (MN) assay and assessed its suitability for immune-correlates analyses and immuno-bridging applications. We analyzed MN and PRNT50 titers measured at baseline and Month 13 in a randomly sampled immunogenicity subset, and at Month 13 in nearly all VCD cases through Month 25. For each serotype, MN and PRNT50 titers showed high correlations, at both baseline and Month 13, with MN yielding a higher frequency of baseline-seronegatives. For both assays, Month 13 titer correlated inversely with VCD risk. Like PRNT50, high Month 13 MN titers were associated with high VE, and estimated VE increased with average Month 13 MN titer. We also studied each assay as a valid surrogate endpoint based on the Prentice criteria, which supported each assay as a valid surrogate for DENV-1 but only partially valid for DENV-2, -3, and -4. In addition, we applied Super-Learner to assess how well demographic, Month 13 MN, and/or Month 13 PRNT50 titers could predict Month 13-25 VCD outcome status; prediction was best when using demographic, MN, and PRNT50 information. We conclude that Month 13 MN titer performs comparably to Month 13 PRNT50 titer as a correlate of risk, correlate of vaccine efficacy, and surrogate endpoint. The MN assay could potentially be used to assess nAb titers in immunogenicity studies, immune-correlates studies, and immuno-bridging applications. Additional research would be needed for assessing the utility of MN titer in correlates analyses of other DENV endpoints and over longer follow-up periods.

Landscapes of binding antibody and T-cell responses to pox-protein HIV vaccines in Thais and South Africans

BACKGROUND

HIV vaccine trials routinely measure multiple vaccine-elicited immune responses to compare regimens and study their potential associations with protection. Here we employ unsupervised learning tools facilitated by a bidirectional power transformation to explore the multivariate binding antibody and T-cell response patterns of immune responses elicited by two pox-protein HIV vaccine regimens. Both regimens utilized a recombinant canarypox vector (ALVAC-HIV) prime and a bivalent recombinant HIV-1 Envelope glycoprotein 120 subunit boost. We hypothesized that within each trial, there were participant subgroups sharing similar immune responses and that their frequencies differed across trials.

METHODS AND FINDINGS

We analyzed data from three trials-RV144 (NCT00223080), HVTN 097 (NCT02109354), and HVTN 100 (NCT02404311), the latter of which was pivotal in advancing the tested pox-protein HIV vaccine regimen to the HVTN 702 Phase 2b/3 efficacy trial. We found that bivariate CD4+ T-cell and anti-V1V2 IgG/IgG3 antibody response patterns were similar by age, sex-at-birth, and body mass index, but differed for the pox-protein clade AE/B alum-adjuvanted regimen studied in RV144 and HVTN 097 (PAE/B/alum) compared to the pox-protein clade C/C MF59-adjuvanted regimen studied in HVTN 100 (PC/MF59). Specifically, more PAE/B/alum recipients had low CD4+ T-cell and high anti-V1V2 IgG/IgG3 responses, and more PC/MF59 recipients had broad responses of both types. Analyses limited to "vaccine-matched" antigens suggested that some of the differences in responses between the regimens could have been due to antigens in the assays that did not match the vaccine immunogens. Our approach was also useful in identifying subgroups with unusually absent or high co-responses across assay types, flagging individuals for further characterization by functional assays. We also found that co-responses of anti-V1V2 IgG/IgG3 and CD4+ T cells had broad variability. As additional immune response assays are standardized and validated, we anticipate our framework will be increasingly valuable for multivariate analysis.

CONCLUSIONS

Our approach can be used to advance vaccine development objectives, including the characterization and comparison of candidate vaccine multivariate immune responses and improved design of studies to identify correlates of protection. For instance, results suggested that HVTN 702 will have adequate power to interrogate immune correlates involving anti-V1V2 IgG/IgG3 and CD4+ T-cell co-readouts, but will have lower power to study anti-gp120/gp140 IgG/IgG3 due to their lower dynamic ranges. The findings also generate hypotheses for future testing in experimental and computational analyses aimed at achieving a mechanistic understanding of vaccine-elicited immune response heterogeneity.

Bridging Efficacy of a Tetravalent Dengue Vaccine from Children/Adolescents to Adults in Highly Endemic Countries Based on Neutralizing Antibody Response

The CYD-TDV vaccine is licensed in multiple endemic countries based on vaccine efficacy (VE) against symptomatic, virologically confirmed dengue demonstrated in two phase 3 trials (CYD14, 2- to 14-year-olds, Asia; CYD15, 9- to 16-year-olds, Latin America). 50% plaque reduction neutralization test (PRNT₅₀) titers at baseline and month 13 (post-vaccination) were associated with VE and may enable bridging VE to adults. Two phase 2 trials of CYD-TDV measured baseline and month 13 PRNT₅₀ titers: CYD22 (9- to 45-year-olds, Vietnam) and CYD47 (18- to 45-year-olds, India). 50% plaque reduction neutralization test distributions were compared between age cohorts, and four versions of an epidemiological bridging method were used to estimate VE against any serotype (dengue virus [DENV]-Any) and against each serotype over 25 months post first vaccination in a hypothetical CYD14 + CYD15 18- to 45-year-old cohort (bridging population 1) and in the actual CYD47 18- to 45-year-old cohort (bridging population 2). Baseline and month 13 geometric mean PRNT₅₀ titers to each serotype were significantly greater in 18- to 45-year-olds than 9- to 16-year-olds for all comparisons. The four methods estimated VE against DENV-Any at 75.3-86.0% (95% CIs spanning 52.5-100%) for bridging population 1 and 68.4-77.5% (95% CIs spanning 42.3-88.5%) for bridging population 2. The vaccine efficacy against serotype 1, 2, 3, and 4 was estimated at 56.9-76.9%, 68.3-85.8%, 91.4-95.0%, and 93.2-100% (bridging population 1) and 44.5-66.9%, 53.2-69.2%, 79.8-92.0%, and 90.6-95.0% (bridging population 2), respectively; thus, CYD-TDV would likely confer improved efficacy in adults than 9- to 16-year-olds. Using the same methods, we predicted VE against hospitalized DENV-Any over 72 months of follow-up, with estimates 59.1-73.5% (95% CIs spanning 40.9-92.2%) for bridging population 1 and 50.9-65.9% (95% CIs spanning 38.1-82.1%) for bridging population 2.

Proportions of CD4 test results indicating advanced HIV disease remain consistently high at primary health care facilities across four high HIV burden countries

Background

Globally, nearly 22 million HIV-infected patients are currently accessing antiretroviral treatment; however, almost one million people living with HIV died of AIDS-related illnesses in 2018. Advanced HIV disease remains a significant issue to curb HIV-related mortality.

Methods

We analyzed 864,389 CD4 testing records collected by 1,016 Alere Pima Analyzers implemented at a variety of facilities, including peripheral facilities, between January 2012 and December 2016 across four countries in sub-Saharan Africa. Routinely collected data and programmatic records were used to analyze the median CD4 counts and proportions of patients with advanced HIV disease by country, facility type, and year.

Results

Median CD4 counts were between 409–444 cells/ul each year since 2012 with a median in 2016 of 444 cells/ul (n = 319,829). The proportion of test results returning CD4 counts above 500 cells/ul has increased slowly each year with 41.8% (95% CI: 41.6–41.9%) of tests having a CD4 count above 500 cells/ul in 2016. Median CD4 counts were similar across facility types. The proportion of test results indicating advanced HIV disease has remained fairly consistent: 19.4% (95% CI: 18.8–20.1%) in 2012 compared to 16.1% (95% CI: 16.0–16.3%) in 2016. The proportion of test results indicating advanced HIV disease annually ranged from 14.5% in Uganda to 29.8% in Cameroon. 6.9% (95% CI: 6.8–7.0%) of test results showed very advanced HIV disease (CD4<100 cells/ul) in 2016.

Conclusions

The proportion of CD4 test results indicating advanced disease was relatively high and consistent across four high HIV burden countries.

Estimation and Inference for Upper Hinge Regression Models

We introduce a new type of threshold regression models called upper hinge models. Under this type of threshold models, there only exists an association between the predictor of interest and the outcome when the predictor is less than some threshold value. Just like hinge models, upper hinge models can be seen as a special case of the more general segmented or two-phase regression models. The importance of studying upper hinge models is that even though they only have one fewer degree of freedom than segmented models, they can be estimated with much greater efficiency. We develop a new fast grid search algorithm to estimate upper hinge linear regression models. The new algorithm reduces the computational complexity of the search algorithm dramatically and renders the existing fast grid search algorithm inadmissible. The fast grid search algorithm makes it feasible to construct bootstrap confidence intervals for upper hinge linear regression models; for upper hinge generalized linear models of non-Gaussian family, we derive asymptotic normality to facilitate construction of model-robust confidence intervals. We perform numerical experiments and illustrate the proposed methods with two real data examples from the ecology literature.

Fc Gamma Receptor Polymorphisms Modulated the Vaccine Effect on HIV-1 Risk in the HVTN 505 HIV Vaccine Trial

HIV Vaccine Trials Network (HVTN) 505 was a phase 2b efficacy trial of a DNA/recombinant adenovirus 5 (rAd5) HIV vaccine regimen. Although the trial was stopped early for lack of overall efficacy, later correlates of risk and sieve analyses generated the hypothesis that the DNA/rAd5 vaccine regimen protected some vaccinees from HIV infection yet enhanced HIV infection risk for others. Here, we assessed whether and how host Fc gamma receptor (FcγR) genetic variations influenced the DNA/rAd5 vaccine regimen's effect on HIV infection risk. We found that vaccine receipt significantly increased HIV acquisition compared with placebo receipt among participants carrying the FCGR2C-TATA haplotype (comprising minor alleles of four FCGR2C single-nucleotide polymorphism [SNP] sites) (hazard ratio [HR] = 9.79, P = 0.035) but not among participants without the haplotype (HR = 0.86, P = 0.67); the interaction of vaccine and haplotype effect was significant (P = 0.034). Similarly, vaccine receipt increased HIV acquisition compared with placebo receipt among participants carrying the FCGR3B-AGA haplotype (comprising minor alleles of the 3 FCGR3B SNPs) (HR = 2.78, P = 0.058) but not among participants without the haplotype (HR = 0.73, P = 0.44); again, the interaction of vaccine and haplotype was significant (P = 0.047). The FCGR3B-AGA haplotype also influenced whether a combined Env-specific CD8+ T-cell polyfunctionality score and IgG response correlated significantly with HIV risk; an FCGR2A SNP and two FCGR2B SNPs influenced whether anti-gp140 antibody-dependent cellular phagocytosis correlated significantly with HIV risk. These results provide further evidence that Fc gamma receptor genetic variations may modulate HIV vaccine effects and immune function after HIV vaccination.IMPORTANCE By analyzing data from the HVTN 505 efficacy trial of a DNA/recombinant adenovirus 5 (rAd5) vaccine regimen, we found that host genetics, specifically Fc gamma receptor genetic variations, influenced whether receiving the DNA/rAd5 regimen was beneficial, neutral, or detrimental to an individual with respect to HIV-1 acquisition risk. Moreover, Fc gamma receptor genetic variations influenced immune responses to the DNA/rAd5 vaccine regimen. Thus, Fc gamma receptor genetic variations should be considered in the analysis of future HIV vaccine trials and the development of HIV vaccines.

Antibody Fc effector functions and IgG3 associate with decreased HIV-1 risk

HVTN 505 is a preventative vaccine efficacy trial testing DNA followed by recombinant adenovirus serotype 5 (rAd5) in circumcised, Ad5-seronegative men and transgendered persons who have sex with men in the United States. Identified immune correlates of lower HIV-1 risk and a virus sieve analysis revealed that, despite lacking overall efficacy, vaccine-elicited responses exerted pressure on infecting HIV-1 viruses. To interrogate the mechanism of the antibody correlate of HIV-1 risk, we examined antigen-specific antibody recruitment of Fcγ receptors (FcγRs), antibody-dependent cellular phagocytosis (ADCP), and the role of anti-envelope (anti-Env) IgG3. In a prespecified immune correlates analysis, antibody-dependent monocyte phagocytosis and antibody binding to FcγRIIa correlated with decreased HIV-1 risk. Follow-up analyses revealed that anti-Env IgG3 breadth correlated with reduced HIV-1 risk, anti-Env IgA negatively modified infection risk by Fc effector functions, and that vaccine recipients with a specific FcγRIIa single-nucleotide polymorphism locus had a stronger correlation with decreased HIV-1 risk when ADCP, Env-FcγRIIa, and IgG3 binding were high. Additionally, FcγRIIa engagement correlated with decreased viral load setpoint in vaccine recipients who acquired HIV-1. These data support a role for vaccine-elicited anti-HIV-1 Env IgG3, antibody engagement of FcRs, and phagocytosis as potential mechanisms for HIV-1 prevention.

Significant Patient Impact Observed Upon Implementation of Point-of-Care Early Infant Diagnosis Technologies in an Observational Study in Malawi

Background

In Malawi in 2014, <20% of human immunodeficiency virus (HIV)–exposed infants received an early infant diagnosis (EID) test in the first 2 months of life and only 30% of HIV-infected children were on antiretroviral therapy (ART). We sought to understand the potential patient impact of improving timely infant diagnosis and treatment initiation through implementation of point-of-care (POC) EID technologies in Malawi.

Methods

In this observational study, POC EID technologies were introduced into routine services at 7 health facilities across Malawi in September 2015. The primary outcome was the proportion of HIV-infected infants initiating ART within 60 days of sample collection in the POC arm compared to the baseline arm with conventional laboratory-based EID testing.

Results

The time from sample collection to result received by the patient decreased significantly from 56 days (interquartile range [IQR], 30–81 days) in the baseline arm to <1 day in the POC arm (P < .001). Of the HIV-infected infants, the time between sample collection and ART initiation was reduced from 38 days (IQR, 30–54 days) in the baseline arm to <1 day (IQR, 0–1 day) in the POC arm (P = .019). Furthermore, the proportion of HIV-infected infants initiated on ART within 60 days of sample collection increased significantly from 41.9% to 91.1% after the introduction of POC (adjusted risk ratio, 2.28; P < .001).

Conclusions

ART initiation rates were significantly improved with the implementation of same-day POC EID testing compared with referred, longer-turnaround laboratory-based testing.

Performance of Cepheid Xpert HIV-1 viral load plasma assay to accurately detect treatment failure

BACKGROUND

Coverage of viral load testing remains low with only half of the patients in need having adequate access. Alternative technologies to high throughput centralized machines can be used to support viral load scale-up; however, clinical performance data are lacking. We conducted a meta-analysis comparing the Cepheid Xpert HIV-1 viral load plasma assay to traditional laboratory-based technologies.

METHODS

Cepheid Xpert HIV-1 and comparator laboratory technology plasma viral load results were provided from 13 of the 19 eligible studies, which accounted for a total of 3790 paired data points. We used random effects models to determine the accuracy and misclassification at various treatment failure thresholds (detectable, 200, 400, 500, 600, 800 and 1000 copies/ml).

RESULTS

Thirty percent of viral load test results were undetectable, while 45% were between detectable and 10 000 copies/ml and the remaining 25% were above 10 000 copies/ml. The median Xpert viral load was 119 copies/ml and the median comparator viral load was 157 copies/ml, while the log10 bias was 0.04 (0.02-0.07). The sensitivity and specificity to detect treatment failure were above 95% at all treatment failure thresholds, except for detectable, at which the sensitivity was 93.33% (95% confidence interval: 88.2-96.3) and specificity was 80.56% (95% CI: 64.6-90.4).

CONCLUSION

The Cepheid Xpert HIV-1 viral load plasma assay results were highly comparable to laboratory-based technologies with limited bias and high sensitivity and specificity to detect treatment failure. Alternative specimen types and technologies that enable decentralized testing services can be considered to expand access to viral load.

Fast Bootstrap Confidence Intervals for Continuous Threshold Linear Regression

Continuous threshold regression is a common type of nonlinear regression that is attractive to many practitioners for its easy interpretability. More widespread adoption of thresh-old regression faces two challenges: (i) the computational complexity of fitting threshold regression models and (ii) obtaining correct coverage of confidence intervals under model misspecification. Both challenges result from the non-smooth and non-convex nature of the threshold regression model likelihood function. In this paper we first show that these two issues together make the ideal approach for making model-robust inference in continuous threshold linear regression an impractical one. The need for a faster way of fitting continuous threshold linear models motivated us to develop a fast grid search method. The new method, based on the simple yet powerful dynamic programming principle, improves the performance by several orders of magnitude.

Point-of-care CD4 technology invalid result rates in public health care settings across five countries

BACKGROUND

Since 2010, point-of-care (POC) CD4 testing platforms have been introduced in both urban and rural settings to expand access to testing by bringing diagnostic services closer to patients. We conducted an analysis of routinely collected CD4 testing data to determine the invalid result rates associated with POC CD4 testing.

METHODS

We analyzed 981,152 CD4 testing records collected from Alere Pima Analyzers between January 2011 and December 2016 across five countries in sub-Saharan Africa. Routinely collected data and programmatic records were used to determine the rate of invalid test results per month, by facility type, and by operator based on cumulative usage during the study period. In addition, frequency of invalid test types and utilization of control beads were assessed.

RESULTS

Across the five countries, 75,530 invalid messages were returned, resulting in an overall invalid result rate of 7.7%. The invalid result rate by country ranged from 6.6% to 11.2%. Invalid result rates were consistent across facility types. Invalid result rates were inversely correlated with operator usage: low volume operators (<50 tests over study period) experienced an invalid result rate of 10.2%, while high volume operators (>500 tests over study period) experienced an invalid result rate of 5.5%. Two invalid result types (exposure position control and reagent control) accounted for nearly 50% of invalid results. Routine data showed that control beads were run on 88.3% of days that the device was used.

CONCLUSIONS

Our analysis found that the rate of invalid results was consistent across all types of health facilities, indicating that decentralization of POC CD4 testing to lower level health facilities did not exhibit high invalid result rates or increase cartridge wastage. Additionally, invalid result rates were inversely correlated to operator usage, with high-volume operators experiencing lower invalid result rates than low-volume operators. POC CD4 testing can, therefore, be performed in decentralized national testing programs; however, adequate training, quality assurance, routine monitoring, and ongoing mentorship should also be implemented for success.

Modification of the Association Between T-Cell Immune Responses and Human Immunodeficiency Virus Type 1 Infection Risk by Vaccine-Induced Antibody Responses in the HVTN 505 Trial

Background

HVTN 505 was a human immunodeficiency virus type 1 (HIV-1) preventive vaccine efficacy trial of a DNA/recombinant adenovirus serotype 5 (rAd5) vaccine regimen. We assessed antibody responses measured 1 month after final vaccination (month 7) as correlates of HIV-1 acquisition risk.

Methods

Binding antibody responses were quantified in serum samples from 25 primary endpoint vaccine cases (diagnosed with HIV-1 infection between month 7 and month 24) and 125 randomly sampled frequency-matched vaccine controls (HIV-1 negative at month 24). We prespecified for a primary analysis tier 6 antibody response biomarkers that measure immunoglobulin G (IgG) and immunoglobulin A (IgA) binding to Env proteins and 2 previously assessed T-cell response biomarkers.

Results

Envelope-specific IgG responses were significantly correlated with decreased HIV-1 risk. Moreover, the interaction of IgG responses and Env-specific CD8+ T-cell polyfunctionality score had a highly significant association with HIV-1 risk after adjustment for multiple comparisons.

Conclusions

Vaccinees with higher levels of Env IgG have significantly decreased HIV-1 risk when CD8+ T-cell responses are low. Moreover, vaccinees with high CD8+ T-cell responses generally have low risk, and those with low CD8+ T-cell and low Env antibody responses have high risk. These findings suggest the critical importance of inducing a robust IgG Env response when the CD8+ T-cell response is low.

Maximum diversity weighting for biomarkers with application in HIV-1 vaccine studies

While studying the association between risk of HIV-1 infection and vaccine-elicited immune responses in preventative HIV-1 vaccine recipients, we encountered a need to combine a collection of biomarkers in an unsupervised fashion with the goal of preserving signal diversity within that collection. Inspired by methods for weighting protein sequences from the biological sequence analysis literature, we propose novel methods for weighting biomarkers, which we call maximum diversity weights. These weights are defined as the weights that maximize measures of signal diversity within a collection of biomarkers. While the optimization problems do not admit analytical solutions, they are convex and hence can be solved efficiently using iterative search algorithms. Through Monte Carlo studies and a real data example from HIV-1 vaccine research, we show that using maximum diversity weights in association studies can lead to an increase in power over other commonly used weights such as uniform weights or principal component-based weights.

Fc Characteristics Mediate Selective Placental Transfer of IgG in HIV-Infected Women

The placental transfer of maternal IgG is critical for infant protection against infectious pathogens. However, factors that modulate the placental transfer of IgG remain largely undefined. HIV-infected women have impaired placental IgG transfer, presenting a unique "disruption model" to define factors that modulate placental IgG transfer. We measured the placental transfer efficiency of maternal HIV and pathogen-specific IgG in US and Malawian HIV-infected mothers and their HIV-exposed uninfected and infected infants. We examined the role of maternal HIV disease progression, infant factors, placental Fc receptor expression, IgG subclass, and glycan signatures and their association with placental IgG transfer efficiency. Maternal IgG characteristics, such as binding to placentally expressed Fc receptors FcγRIIa and FcγRIIIa, and Fc region glycan profiles were associated with placental IgG transfer efficiency. Our findings suggest that Fc region characteristics modulate the selective placental transfer of IgG, with implications for maternal vaccine design and infant health.