The IMpact of Vertical HIV infection on child and Adolescent SKeletal development in Harare, Zimbabwe (IMVASK Study): A protocol for a prospective cohort study

The scale-up of antiretroviral therapy (ART) across sub-Saharan Africa (SSA) has reduced mortality so that increasing numbers of children with HIV (CWH) are surviving to adolescence. However, they experience a range of morbidities due to chronic HIV infection and its treatment. Impaired linear growth (stunting) is a common manifestation, affecting up to 50% of children. However, the effect of HIV on bone and muscle development during adolescent growth is not well characterised. Given the close link between pubertal timing and musculoskeletal development, any impairments in adolescence are likely to impact on future adult musculoskeletal health. We hypothesise that bone and muscle mass accrual in CWH is reduced, putting them at risk of reduced bone mineral density (BMD) and muscle function and increasing fracture risk. This study aims to determine the impact of HIV on BMD and muscle function in peripubertal children on ART in Zimbabwe. Methods and analysis Children with (n=300) and without HIV (n=300), aged 8-16 years, established on ART, will be recruited into a frequency-matched prospective cohort study and compared. Musculoskeletal assessments including dual-energy X-ray absorptiometry, peripheral quantitative computed tomography, grip strength and standing long jump will be conducted at baseline and after 1 year. Linear regression will be used to estimate mean size-adjusted bone density and Z-scores by HIV status (ie, total-body less-head bone mineral content for lean mass adjusted for height and lumbar spine bone mineral apparent density. The prevalence of low size-adjusted BMD (ie, Z-scores <-2) will also be determined. Ethics and dissemination Ethical approval for this study has been granted by the Medical Research Council of Zimbabwe and the London School of Hygiene and Tropical Medicine Ethics Committee. Baseline and longitudinal analyses will be published in peer-reviewed journals and disseminated to research communities. This study aims to determine the impact of HIV infection on size-adjusted bone density in peripubertal children aged 8–16 years established on ART. In addition, this prospective study aims: We hypothesise that HIV infection adversely affects skeletal development, such that CWH, despite ART, accrue less bone mass and strength and have reduced muscle function during skeletal development. CWH aged 8–16 years and established on ART (n=300) and a comparison group of children without HIV, frequency-matched for age and sex (n=300) will be recruited into a prospective cohort study. Detailed musculoskeletal assessments will be conducted at baseline and after 1 year. Parirenyatwa and Harare Hospital are the largest public-sector referral hospitals in Harare.40 41 The paediatric HIV clinics at both hospitals provide HIV care to more than 2000 children. Although HIV care is increasingly decentralised to primary care level across the country, most children in Harare continue to receive care within HIV clinics in secondary healthcare facilities. Parirenyatwa hospital has a well-functioning radiology department which houses the University of Zimbabwe DXA and pQCT research unit and has access to private radiology services in the surrounding area. The hospital catchment areas have over 116 primary and 42 secondary government schools with an estimated 157 962 children enrolled.42 School attendance in Harare province is high and does not differ by HIV status, with 96% of children under 18 years attending school.43 Inclusion criteria: age 8–16 years (includes prepubertal and peripubertal children), living in Harare, and in CWH only if: Children with perinatally acquired HIV will be included in this study. Perinatally acquired HIV will be defined based on Zimbabwean criteria, that is, self-report of no sexual debut or blood transfusions, a history of natural sibling or maternal HIV and characteristic clinical features of longstanding HIV. Children with horizontal infection will also be included in the study. Exclusion criteria: acute illness (requiring immediate hospitalisation) and lack of consent. Detailed information on all the above comorbidities will be collected using the main study questionnaire in the clinical history section. This information will be collected for both children with and without HIV. Comorbidities will not be used as the basis of excluding children from the study. However, for the purposes of deriving normative DXA data, those with severe bone disease will be excluded at the analysis stage. Systematic quota-based sampling by age and sex will be used to recruit 300 children from Parirenyatwa and Harare Hospital HIV clinics. Participants will be recruited sequentially as they attend clinic such that 50 male and 50 female will be chosen for each of three age strata: 8–10.99, 11–13.99 and 14–16.99 years. A maximum of five participants will be enrolled on each day for logistical reasons. The total number of children approached each day will be recorded, irrespective of whether they are subsequently eligible or enrolled to determine the sampling fraction. Written consent will be obtained from children and their guardians. Study processes and procedures will be clearly explained to children and their guardians and they will be given the option to accept or decline to take part in the research. It will also be explained that they are allowed to withdraw from the study at any time, for any reason, without affecting the care they receive from the clinic. Three hundred CWH will be randomly sampled from six government primary and secondary schools in the same catchment area as Parirenyatwa and Harare Hospitals. Younger children (8–12 years) will be selected from primary schools and older children (13–16 years) from secondary schools, with 13 year olds coming from both primary and secondary schools. The number of children selected from each school will be proportional to school size, thereby giving each child equal probability of being sampled. A random number sequence will be generated, and school registers will be used to select participants of similar age and sex as the children with HIV using the same quota-based approach of 50 male and 50 female in each of the three age strata. Guardians of selected school children will be invited to the study clinic to complete the consent process. Consenting participants will have a diagnostic HIV test as part of their assessment. Those testing HIV positive (anticipated to be approximately 2%–3%44) will be referred for HIV care. An interviewer-administered questionnaire together with hand-held medical records will be used to collect sociodemographic details and clinical history including age, sex, school attendance, orphan status, guardianship, history of fractures with mechanism of trauma, steroid use, smoking, alcohol, recreational drugs, family history of musculoskeletal disease, comorbidities, physical activity, diet and nutrition, and sun exposure. Where possible, validated instruments adapted for the local context will be used. For example, the International Physical Activity Questionnaire (IPAQ)45 validated in multiple countries including South Africa and will be used to assess physical activity as multiples of the resting metabolic rate (MET) in MET-minutes. Diet and nutrition will be assessed using a tool we developed for the Zimbabwean context based on a validated dietary diversity and food frequency tool from India and Malawi46 and international guidelines applicable to SSA.47 The tool quantifies vitamin D supplementation and sunlight exposure and has been adapted to reflect the Zimbabwean context where fortification of oils and margarine with vitamin D is mandated by the government and specific vitamin D–rich foods such as kapenta fish are found. A standardised musculoskeletal examination will be conducted using the validated paediatric Gait, Arms, Legs and Spine (pGALS) examination.48 Additional clinical assessments will be carried out using standardised protocols and calibrated equipment. Anthropometry measurements will include standing and sitting height, arm span and mid-upper arm circumference. Height will be measured to the nearest 0.1 cm, by two separate readers using calibrated Seca 213 stadiometers. If the two height measurements differ by more than 0.5 cm, a third reading will be taken.49 Weight will be measured to the nearest 0.1 kg using calibrated Seca 875 scales. Tanner pubertal staging will be carried out using a standardised protocol with an orchidometer to assess testicular volume in male participants.50 Muscle function will be assessed in the upper and lower limbs by grip strength dynamometry and standing long jump, respectively. Hand grip strength will be measured using a Jamar hydraulic hand-held dynamometer (Patterson Medical, UK) to the nearest 0.1 kg. Participants will be seated with the shoulder at 0° to 10°, the elbow at 90° of flexion and the forearm positioned neutrally. Three measurements will be taken from each hand in alternation and the highest measurement chosen. The standing long jump distance will be taken from the best of three correctly performed attempts to the nearest 0.1 cm, measuring the distance from the take-off line to the heel. DXA scans will be performed by two trained radiographers using a Hologic QDR Wi densitometer with Apex software V.4.5. Measurements will be taken from the lumbar spine, hip and total body. Fat and muscle mass will also be acquired; muscle mass is the fat-free mass measurement from DXA. DXA scans will be repeated in a subgroup (n=20) of participants to determine reproducibility. pQCT measurements of the non-dominant tibia will be taken using a Stratec XCT-2000 scanner (Stratec, Pforzheim, Germany) software V.6.20. Measurements of the non-dominant tibia will be taken at three sites at 4%, 38% and 66% of the tibial length, measured from the medial malleolus to the medial tibial plateau. Daily quality control will be performed by scanning the manufacturer-provided lumbar spine phantom for DXA and tibia phantom for pQCT. A radiograph of the non-dominant hand and wrist will be taken and used to quantify bone age using the Greulich and Pyle (G&P) atlas and the Tanner Whitehouse 3 (TW3) method. For intraobserver reliability, 10% of the radiographs will be randomly selected and rescored by the same operator after 1 week. For interobserver reliability, a different set of 10% of the radiographs will be rescored by a different expert. The estimated bone age will then be compared with the calculated chronological age. A fasting blood sample (up to 15 mL) will be collected. HIV markers (CD4 count and viral load) will be tested in CWH only. CD4 cell count will be measured using an Alere PIMA CD4 machine (Waltham, Massachusetts, USA). HIV viral load will be measured using the GeneXpert HIV-1 viral load platform (Cepheid, Sunnyvale, California, USA). The remaining blood plasma will be biobanked to enable future measurement of bone biochemistry. After removing the plasma, peripheral blood mononuclear cells will be isolated and cryopreserved. DNA will also be extracted using a manual method and stored for future genetic studies. All study measurements, with the exception of DNA extraction, will be repeated after 1 year. Participants will be recalled exactly 1 year after their first DXA scan. The aim is to perform all scans within a 4-week window period. Contact will be maintained with participants via regular phone calls and text messaging to minimise loss-to-follow-up. The schedule of study procedures is summarised in table 1. Summary of study measurements to be quantified at baseline and follow-up *Details of treatment and comorbidities will be confirmed by patient-held medical records where available. †Denotes assessments to be carried out in HIV-infected participants only. ‡Energy requirements defined in METS (multiples of the resting metabolic rate that give a score in MET-minutes). §Nutritional indicator to include composite information from history (usual diet last month, sun exposure–vitamin D status) and clinical examination (MUAC). Similar methods have been used in other low-income contexts.46 ¶Age-specific and sex-specific Z-scores for (1) anthropometric measures will be determined using the WHO child growth standards.59 (2) Hand grip strength will be determined with reference to the uninfected comparison group and European normative data.60 (3) Jumping distance will be determined using normative data from South Africa.61 (4) Low BMD will be determined with reference to published paediatric Hologic DXA reference databases for LS BMAD and TBLH BMCLBM Z-scores.36 **Standing long jump; the longest distance after two attempts will be recorded. ††Pregnancy urine dipstick in female participants prior to DXA if uncertain pregnancy status. ‡‡Tests to be carried out on stored blood when further funding is secured. CSA, cross-sectional area; CSMI, cross-sectional moment of inertia; DXA, dual-energy X-ray absorptiometry; LS BMAD, lumbar spine bone mineral apparent density; PMI, polar moment of inertia; SSI, Strength Strain Index; TBLH BMCLBM, total-body less-head bone mineral content for lean mass adjusted for height. The primary study outcomes are as follows: Secondary study outcomes are as follows: The sample size was calculated to detect differences in DXA-measured mean size-adjusted bone BMD Z-scores between children with and without HIV. This study will have 80% power (α 0.05) to detect a 0.23 Z-score difference between 300 HIV-infected and 300 uninfected children, assuming an SD of 1.3. As there were no published studies from low income countries, estimates of the expected difference were taken from a US study of children with HIV aged 7–15 years.10 In addition, our study will have 80% power to detect a 4.8% difference in the prevalence low size-adjusted BMD between the two groups. This is a smaller prevalence difference than that detected by the most conservative prevalence estimate of low BMD of 7% from three studies in high-income and middle-income countries.10–12 For continuous variables that are normally distributed, the mean and SD will be presented. For skewed continuous variables, the median and IQR will be presented. Categorical variables will be summarised as frequencies and percentages. The distribution of demographic and clinical variables will be compared between CWH and without HIV using independent sample t-tests for means, Wilcoxon rank-sum test for medians and χ2 tests for proportions. Baseline mean TBLH BMCLBM and LS BMAD Z-scores and the prevalence of low TBLH BMCLBM and LS BMAD Z-score will be compared between CWH and without HIV. Among CWH, the association between a priori defined risk factors (ART duration, ART type, proportion of life on treatment, age at ART initiation, CD4 count, viral load, bone age, pubertal stage, nutrition, socioeconomic status and orphanhood) against size-adjusted BMD will be examined using multivariable linear regression (Z-score as a continuous variable) and multivariable logistic regression (as defined by the Z-score cut-off of <−2). Successful antiretroviral therapy will be defined as a viral load of less than 1000 copies/mL. Paired sample t-test or non-parametric Wilcoxon test will be used to assess for differences in TBLH BMC and LS BMAD on children with and without HIV between baseline and follow-up. Multivariable linear regression will be used to analyse the mean percentage change in TBLH BMCLBM (g) and LS BMAD (g/cm3) between children with and without HIV. Models will be adjusted for physical activity, calcium and vitamin D intake. Interaction between the effects of pubertal stage (bone age) and HIV on change in TBLH BMCLBM and LS BMAD will be investigated to see if differences in bone density become more or less pronounced through puberty, that is, whether catch-up growth is possible, see figure 2. The regression coefficient (β) for percentage change in size-adjusted bone mass may suggest either no growth impairment (figure 2A), delayed puberty while maintaining the same growth trajectory (figure 2B) or delayed puberty with a reduced growth trajectory (figure 2C) in CWH. If β is markedly more positive in CWH, this suggests that catch-up growth may be possible (figure 2D). Pubertal delay in this study will be defined as the lack of the initial signs of puberty (Tanner stage 2) at an age that is more than 2 SD beyond the population mean51 and as chronological age minus bone age >2 years.52 Data for total body and lumbar spine DXA, tibial pQCT, hand grip strength and standing long jump in CWH will be analysed with reference to the comparator group of children without HIV. Hypothesised growth scenarios to be assessed as interactions between pubertal stage and HIV status on change in bone mass. For the purposes of normative data derivation, children without HIV who have any diagnosis or evidence of muscle or bone disease will be excluded. Then outliers with bone density, hand grip strength or standing long jump data beyond 2 SD from the mean will have their case record reviewed to exclude cases with underlying bone or muscle pathology. The remaining population will be used to generate normative references ranges for these quantitative traits. Data collection, management and storage will be governed by standard operating procedures and will follow the principles of Good Clinical Practice (GCP). Data will be captured using hand-held tablets for the questionnaires. Paper forms will be available in case of failure of electronic data entry. Microsoft Access will be used as the main back-end database as it allows programming of quality control checks and conditional data validation. GCP compliant audit trail modules will be incorporated into the databases and reports of aggregated data will be reviewed on a monthly basis. In order to assure data quality and consistency, all staff will receive regular training and regular quality checks will be conducted. Paper records will be stored for 8 years after the completion of research in secure, locked storage facilities. Field staff will download data to the central database, which is backed up onto an encrypted external hard drive daily, and to additional off-site and secure cloud back-up. The off-site back-up copies will be stored through the London School of Hygiene and Tropical Medicine (LSHTM) Research Data Management Support Service that has an established data repository. In order to preserve the long-term value of these data, it will be stored backed-up here indefinitely. Anonymised research data will be made available for sharing through the open access data repository established by the LSHTM Data Management Support Service at the time of publication. This will allow other research groups to request access to study data and tools. Information on other researchers’ data will be included in every study publication. While patients were not directly involved in the design and conduct of the study, feedback from patient experiences in the study will be used to inform planned public engagement activities, which include science fairs, conducted by the research team at schools from where participants were recruited. Recruitment to this study began in May 2018 and is planned until August 2019. Study follow-up will run from May 2019 to August 2020.