Inpatient Point-of-Care HIV Early Infant Diagnosis in Mozambique to Improve Case Identification and Linkage to Antiretroviral Therapy

DISCUSSION

This study shows that inpatient POC EID is both feasible and effective. Not only did the placement of POC technology in these hospital settings increase overall testing volume and newly identify a high number of children with HIV-positive tests, but the immediate onsite POC test results also permitted much improved linkage to ART for these children compared with the pre-POC period.

A highly significant 97% increase in testing volume per month occurred after allocation of POC, which we believe is best explained by clinicians being more likely to order EID testing when timely results were available. Improved inpatient PITC at both hospitals during the time period of the study could have contributed to more newly identified HEIs, but PITC testing data were not included in this study and the large majority of EID tests are performed in infants already known to be HIV exposed. No dramatic changes in maternal antenatal HIV prevalence, annual admission volumes, or EID eligibility criteria occurred during the time period of the study that would otherwise explain this large increase in testing volume. Yet, despite the large increase in testing volume, the POC platforms still may have been underutilized. This circumstance is of particular concern at HCM, which is in a higher-HIV prevalence province and has approximately 60% more admissions per month, but had fewer inpatient POC tests performed per month than did HCB (16.5 versus 18.3).¹⁹ Increased POC EID testing is needed in wards that admit HEIs 12–18 months of age, which represented only 36.6% of the test requests in this study. This finding is important in light of new EID algorithm recommendations from the World Health Organization that call for virologic testing in HEIs >9 months of age given the possibility of false-negative rapid HIV test results.^11,20–22

Concerningly, a study of pediatric inpatient PITC from wards in Mozambique reported that coverage rates are low.¹³ Routine opt-out rapid antibody testing of breastfeeding mothers is a crucial first step to newly identify HEIs who need EID testing, and this step can help improve POC EID platform utilization. The same study also showed that hospitals that relied on DBSs for EID often did not test or retest exposed infants and presumptive HIV diagnosis was rarely made.¹³ POC platforms bypass the inherent delays that come with DBS-based EID; eliminate the need for inpatient presumptive HIV diagnosis, which has been consistently underutilized; and have the potential to strengthen pediatric inpatient PITC and improve case detection. Inpatient POC EID can also serve as an important backstop to traditional outpatient HEI-clinic EID, as 37.5% of the infants with positive POC tests had a previous negative outpatient PCR and were presumably infected during breastfeeding.

The 29.7% overall positivity rate for POC EID from this study is significantly higher than other results reported from the region. A POC EID evaluation from 8 African countries reported a 15.2% prevalence in pediatric inpatients (hospitalized children represented only 2.9% of the total cohort), and a study from Malawi, where 48.9% of the infants tested with POC came from inpatient wards, had an overall positivity rate of 5.7% (no inpatient/outpatient disaggregation provided).^5,7

The EID positivity rate was even higher before POC implementation, at 45.5%. Although some of this difference could be explained by improved PMTCT coverage in the later POC period, it is also likely that DBS testing was only performed in HEIs with more advanced signs and symptoms of HIV infection.¹ Not all admitted HEIs need repeat virologic testing with POC, but the ease of use and timely results seem to lower the clinical threshold for testing to include more subtle early signs of infection, thereby allowing for more timely diagnosis of HIV with better treatment outcomes.

The PMTCT program in Mozambique continues to struggle to reduce vertical transmission, with an estimated rate of 15% (much of which occurs via breastfeeding), challenges with maternal ART adherence and retention, and high rates of maternal seroconversion during pregnancy and lactation.^23–26 Undiagnosed infants will often become ill and require hospitalization, and this study showed the high toll of such late presentation in terms of inpatient mortality at 28.2%. This rate was higher than the 22% mortality rate reported in a recent pediatric inpatient study from Kenya that included infants and older children, and it highlights the importance of strengthening serial maternal HIV testing throughout pregnancy and the breastfeeding period, as well as outpatient EID, so diagnoses can be made before children develop advanced disease requiring hospitalization.²⁷

Confirmed ART initiation for infants with positive inpatient EID tests was 64% higher in the POC period, facilitated by same-day actionable results. Despite previous efforts to prioritize central laboratory testing of DBS samples from inpatient wards, the turnaround times for conventional DNA PCR (36-day average for the 11 largest hospitals in Mozambique in 2016) meant that most hospitalized patients had been discharged before their results were returned, despite being expedited.¹² Furthermore, posthospitalization follow-up often occurs at primary health care centers closer to patients' homes, and the resources needed for routine active tracing of infected infants via phone or home visits are generally lacking. Consequently, these conventional results never reached the children's caregivers, contributing to reduced linkage to ART. This study shows that such hurdles to timely ART initiation for infants with HIV infection diagnosed during hospitalization can be addressed with the use of POC EID.

With the exclusion of infants who died during hospitalization prior to ART initiation, 61.2% of those with positive tests initiated as inpatients. Our data set did not allow for analysis of why some infants did not initiate ART as inpatients, but anecdotally the most common reason was inpatient tuberculosis diagnosis with guidelines recommending 2 weeks of treatment before ART should be started. An additional 29.8% of HIV-infected infants who were discharged without ART were confirmed to have initiated as outpatients. These numbers compare unfavorably to other POC studies, which reported ART linkage rates of 86.3% and 91%, but did not disaggregate inpatient versus outpatient timing.^5,7 A large difference was found in confirmed ART post discharge in patients who initiated as inpatients versus those who did not (66.1% vs. 29.8%), suggesting that an ideal model would promote ART initiation during admission to the extent possible. A recent clinical trial showed no early mortality benefit to urgent inpatient ART initiation in children, but the findings from this study suggest that initiation after stabilization but before discharge may help improve access to and retention on ART.²⁷ A clear need exists to improve active follow-up of infants diagnosed during hospitalization at these sites to ensure successful linkage to ongoing ART.

The results from this study suggest that POC EID can further improve access to EID and pediatric ART by broadening the considerations for potential placement scenarios. Our findings suggest there is clear benefit to deploying POC testing to hospitals with subsequent establishment of context-appropriate ART initiation models that readily decentralize back to primary care. However, this consideration for further deployment of POC testing to inpatient settings does not need to be limited to large pediatric hospitals; the dynamics of inpatient reality (late presentation, incomplete EID testing history, admission duration shorter than conventional DNA PCR turnaround times) are applicable to smaller wards throughout the country whose patients would also benefit from inpatient POC EID. The ability to now multiplex on POC instruments, such as the HIV viral load assays that are now also being performed at HCM and HCB on the Abbott mPIMA analyzer used in this study, the advent of birth EID testing opportunities, and combined inpatient/outpatient testing at sites with both EID clinics and pediatric wards can generate testing volumes that justify deployment to settings such as district hospitals that previously may not have met POC allocation criteria. Indeed, there remain deployment opportunities to further reach HEIs across the health care system, and POC consistently proves to be an effective and feasible solution—one that will be needed to help reach national and global targets.