Risk factors for pneumococcal carriage in adults living with HIV on antiretroviral therapy in the infant pneumococcal vaccine era in Malawi

27

28 Adults living with HIV (ALWHIV) on antiretroviral therapy (ART) are at high risk of pneumococcal 29 carriage and disease. To help evaluate carriage risk in African ALWHIV in the infant pneumococcal 30 conjugate vaccination era, we assessed association between carriage and potential risk factors. 31 32

33 Nasopharyngeal swabs were collected from adults aged 18-40 years attending an ART clinic during 34 rolling, cross-sectional surveys in Blantyre, Malawi between 2015-2019. We fitted generalised 35 additive models to estimate the risk of sex, social economic status (SES), living with a child <5y, and 36 ART duration on carriage. 37 38

39 Of 2,067 adults, median age was 33y (range 28-37), 1,427 (69.0%) were females, 1,087 (61.4%) were 40 in low-middle socio-economic-status (SES), 910 (44.0%) were living with a child <5y, and median 41 ART duration was 3.0 years (range 0.004-17). We estimated 38.2% and 60.6% reductions in overall 42 and vaccine-serotype carriage prevalence. Overall carriage was associated with low SES, living with a 43 child <5y and shorter duration on ART. By contrast, vaccine-type carriage was associated with living 44 without a child <5y and male sex. 45 46

47 Despite temporal reductions in overall and vaccine-serotype carriage, there is evidence of incomplete 48 VT indirect protection. A targeted-vaccination campaign should be considered for ALWHIV, along 49 with other public health measures to further reduce vaccine-serotype carriage and therefore disease. 50 51 Key words: Pneumococcal carriage, pneumococcal conjugate vaccine, human immunodeficiency 52 virus, herd immunity, antiretroviral, risk factors, Malawi 53 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 3

54 Streptococcus pneumoniae (the pneumococcus) is a common coloniser of the human nasopharynx, 55 particularly in young children and populations with human immunodeficiency virus (HIV) 1. 56 Pneumococcal colonisation is a necessary prerequisite for transmission and the development of 57 disease, including otitis media, sinusitis, pneumonia, meningitis, and bacteraemia 2. The 58 pneumococcus is associated with a large burden of disease in adults living with HIV (ALWHIV) 59 compared to adults without HIV 3–5. Adult HIV prevalence remains high (>10%) in many sub-Saharan 60 African countries, with Malawi reporting a national prevalence of 10.6% 6–8. The use of antiretroviral 61 therapy (ART) has substantially increased survival and reduced the incidence of invasive 62 pneumococcal disease (IPD) 9. However, despite more than 85% of ALWHIV in Malawi receiving 63 ART 10–12, ALWHIV remain at greater risk of IPD than adults without HIV 3. 64 65 Pneumococcal conjugate vaccines (PCVs) are widely used in infant schedules in low- and middle-66 income countries (LMICs), generally targeting the most commonly invasive serotypes in this age 67 group 1. To date, in contrast to high-income settings, immunisation of vulnerable adults with 68 pneumococcal vaccines has not been adopted in most LMICs 13. In November 2011, Malawi 69 introduced the 13-valent PCV (PCV13) into the national Expanded Program on Immunisation (EPI) 70 using a three-primary-dose schedule without booster (3+0; one dose at 6, 10 and 14 weeks of age). 71 Despite nearly 10 years of >80% PCV13 three-dose coverage among age-eligible children, there is 72 evidence of a sub-optimal reduction in both vaccine-serotype (VT) carriage prevalence and VT-IPD 73 incidence in children and ALWHIV in Malawi 14–16. Similar evidence of residual VT carriage 74 prevalence is also reported in the Gambia and Mozambique after 5 and 2 years of implementation, 75 respectively, 17 despite both countries also reporting >80% PCV three-dose coverage under a 3+0 76 vaccine schedule 18,19. There is increasing evidence that the indirect protection (i.e. herd immunity) 77 offered by an infant PCV against VT carriage, especially in ALWHIV, is sub-optimal 14,18. 78 79 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 4 The most effective strategy to reduce residual VT-IPD burden in ALWHIV depends on the factors 80 shaping VT carriage and disease risk. Before introducing infant PCV in Malawi and South Africa, risk 81 factors for IPD in ALWHIV included younger age, female sex, cotrimoxazole resistance, underlying 82 medical conditions and living in a densely populated area 3,4. On the other hand, risk factors for 83 carriage of any pneumococcal serotype included exposure to infants exposed to HIV 20,21, low socio-84 economic status (SES), high density living in housing with inadequate ventilation and with intense 85 social contacts 22–25. Moreover, among Malawian ALWHIV, the overall (VT and non-VT [NVT]) 86 carriage prevalence was found to be higher in those on ART than not on ART 26,27. 87 88 In the PCV era, however, there are important information gaps in our understanding of the relative 89 importance of of key factors for pneumococcal carriage and disease among ALWHIV. These include 90 duration on ART, PCV vaccination status among children in the household, and SES. Following our 91 recent data from Blantyre, Malawi showing high residual VT carriage and its determinants in PCV13-92 vaccinated and unvaccinated children and in ALWHIV 15,16, we extend the analysis to identify age- 93 and time-dependent risk factors for pneumococcal carriage in ALWHIV on ART using generalised 94 additive modelling. 95 96

97 Study design 98 Blantyre is located in the southern region of Malawi spanning 2,025 km2, with an urban (population 99 density 3,334/km2) and rural (253/km2) population of approximately 800,000 and 451,000 people, 100 respectively. 28. As described elsewhere in more detail15, rolling, prospective cross-sectional 101 pneumococcal nasopharyngeal (NP) carriage surveys were conducted between 29 June 2015 and 9 102 August 2019 in Blantyre to investigate the change over time of pneumococcal colonisation in 103 ALWHIV on ART. The majority (98.6%) of sampled individuals were on a first line ART regimen 104 containing either i) Zidovudine, Lamivudine and Efavirenz, ii) Tenofovir, Lamivudine and Efavirenz 105 or iii) Tenofovir, Lamivudine and Nevirapine 29. Eight pneumococcal carriage surveys (each 106 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 5 approximately 6 months in duration) were conducted from 3.6 to 7.9 years after infant PCV13 107

into the EPI schedule. ALWHIV aged 18-40 years were recruited from the Queen 108 Elizabeth Central Hospital (QECH) ART clinic in Blantyre using a systematic sampling approach. 109 Exclusion from the study included being currently on treatment for tuberculosis, hospitalisation within 110 two weeks of recruitment and previously enrolled in the survey. 111 112 Nasopharyngeal sample collection and processing 113 An NP swab sample was collected from each participant and processed at the Malawi-Liverpool-114 Wellcome Programme laboratory, co-located to QECH, to ascertain the presence of pneumococci. 115 Samples were collected and processed according to World Health Organisation guidelines 30. 116 Serotyping was done using latex agglutination, based on picking a single colony, to identify serotypes 117 targeted by PCV13 (1, 3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F, 23F). Non-typeable and NVT 118 isolates were both classified as NVT. Pneumococcal carriage dynamics were further evaluated using 119 DNA microarray techniques, a technique which, in the case of co-carriage of multiple pneumococcal 120 serotypes, differentiates all individual serotypes and reports relative abundance of each serotype in 121 carriage 31,32. Microarray was implemented only in surveys 1 through 4 and with samples having 122 latex-confirmed pneumococcal carriage. Further details of sample processing has been reported earlier 123 15,32. 124 125 Data collection and analysis 126 Participant data collected at recruitment included age, sex, cohabitation with a child <5y (Yes/No), 127 social economic status (SES), duration of ART use, CD4+ T-cell count, current ART regimen and 128 cotrimoxazole use. A multiple imputation random forest-based method, using MissForest R package 129 34, was conducted to impute 1 (0.0005%), 297 (14.4%), and 537 (26.0%) missing data points on 130 cohabitation with a child <5y, SES, and duration of ART use, respectively. Though reported in the 131 descriptive analysis, CD4+ cell count was excluded from model-based analyses because 46.0% of its 132 data points were missing, a proportion above the acceptable standard threshold for imputation 34. 133 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 6 Duration on ART was not used as a continuous variable because of data sparsity in age- or time-134 stratified analyses but was categorised as short (<3 years) or long (≥3 years) duration based on (1) a 135 previous study in rural Malawi which showed strong evidence of high pneumococcal carriage during 136 the first 2 years of ART use 26, and (2) the median value of ART duration in this study. 137 138 Individual fitted carriage prevalence estimates were categorised into 18-24y, 25-29y, 30-34y, and 35-139 40y age groups reflecting their distinct IPD incidence 3. Time was stratified into year of survey 140 initiation (2015, 2016, 2017, 2018 and 2019). In Malawi, pneumococcal carriage prevalence is usually 141 higher in the cold (May-August) and hot (September-November) seasons as compared to the rainy 142 season (December-April) 26. Thus, seasonality in carriage was captured using an indicator variable 143 with values ranging from January to December based on the month of NP sample collection. 144 145 Generalised additive modelling framework 146 We used a generalised additive modelling (GAM) framework to fit to age- and time-specific 147 trajectories of pneumococcal carriage, and allow flexibility in capturing nonlinear carriage dynamics. 148 In brief, we used penalised B splines (P-splines) for the age and time spline smoothers to avoid knot 149 selections which usually introduce under- and over-fitting biases when trading-off model fit to the 150 data and the smoothness of the curve 35. A penalized log-likelihood maximization was used to fit a 151 non-parametric binomial model with complementary log-log link function defined by log-hazard of 152 carriage as a function of the risk factors and a spline in age and time. No time-series autocorrelation 153 structure was included in the model fits because ALWHIV were independently sampled without 154 replacement and with no evidence to suggest strong autocorrelation across time. Thus, the rolling, 155 prospective cross-sectional carriage samples and model residuals were assumed to be serially 156 independent. 157 158 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 7 Age- and time-dependent carriage prevalence estimation 159 We modelled age- and time-dependent carriage trajectories separately for overall (VT+NVT) and VT 160 carriage as outcome variables for a set of potential risk factors. Due to reported poor immunogenicity 161 and effectiveness of PCV13 against serotype 3 36,37, we also modelled VT carriage without serotype 3 162 (VT-st3) to explore changes in carriage prevalence. A model with age and time smoothers and 163 potential risk factors including sex, seasonality, duration on ART, cohabitation with a child <5 years 164 old, and SES were fitted to the carriage data to estimate the overall or VT carriage prevalence and risk 165 factor-specific effects on carriage prevalence dynamics. 166 167 A ‘gam’ function in the ‘mgcv’ R package facilitated model fitting 38, based on a model formulated as 168 /g1859/g3435/g1842 /g4666 /g1851 /g3036/g34041 | /g1853 /g3036,/g1872 /g3036/g4667 /g3439/g3404/g1859 /g3435 /g2024 /g4666 /g1853 /g3036,/g1872 /g3036/g4667 /g3439/g3404/g2015 /g4666 /g1853 /g3036,/g1872 /g3036/g4667 , where /g1851 /g3036 is a binomial outcome on whether an 169 individual /g1861 is carrying pneumococcus (1) or not (0); /g1859 is the complementary log-log link function; 170 /g2024 /g4666 /g1853 /g3036,/g1872 /g3036/g4667 is the carriage prevalence estimate for individuals of age (/g1853 /g3036) at time (/g1872 /g3036); /g2015/g4666/g1853 /g3036,/g1872 /g3036/g4667 is a 171 nonparametric linear predictor as function of individual age and time, and a set of risk factors. The 172 linear predictor on a predictor scale is further expanded using the equation 173 /g2015 /g4666 /g1853 /g3036,/g1872 /g3036/g4667 /g3404/g2010 /g2868/g3397 ∑ /g2010 /g3038/g1833 /g3036 /g3038/g3397/g1872 /g1857 /g4666 /g1853 /g3036/g4667 /g3397/g1872 /g1857 /g4666 /g1872 /g3036/g4667 , where /g2010 /g2868 is a model intercept, /g1833 /g3036 refers to individual risk 174 factor category, /g2010 /g3038 is the risk factor coefficient, /g1872/g1857 /g4666 /g1853 /g3036/g4667 and /g1872/g1857 /g4666 /g1872 /g3036/g4667 denote tensor product P-spline of 175 predictor age (/g1853 /g3036) and time (/g1872 /g3036). 176 177 The relative difference in carriage prevalence was computed by subtracting the GAM carriage 178 prevalence estimate for each age or time category from the reference category and then dividing the 179 difference by the reference category and then multiplying by 100%. The 95% confidence interval 180 (95%CI) of the relative difference was estimated using 181 /g4666 /g2025/g33971 /g4667 /g1499/g4670 1 ± 1.96 /g1499 /g3493 /g2012 /g2869 /g2870/g3397/g2012 /g2870 /g2870/g3398 /g46661.96 /g2870/g4667/g1499/g2012 /g2869 /g2870/g1499/g2012 /g2870 /g2870/g46661 /g3398 1.96 /g1499 /g2012 /g2869 /g2870/g4667/g3415 /g4671/g33981 , where /g2025 is the relative 182 difference, /g2012 /g2869 and /g2012 /g2870 are the coefficient variations of the reference and comparator categories, 183 respectively, and coefficient variation being standard deviation divided by the observed mean 39. 184 GAMs with and without interactions between age group or time and each independent risk factor on 185 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 8 the overall and VT carriage prevalence were fitted and compared using Akaike information criterion 186 (AIC) , and results of these tests are numerically presented in S1 Table. Given the model 187 complexities, sensitivity analyses assesed factors that may affect carriage estimates which included 188 the impact on carriage of individual age group or survey, serotyping method, carriage autocorrection, 189 model formulation and spline type. Detailed sensitivity methods and results are presented in 190 Supplementary Material (S1 Text and S2 Text). Analyses were conducted in R v4.1.1 40, with 191 statistical significance set at p<0.05, and the code is publicly shared via GitHub 41. 192 193 Ethical approval 194 Ethical approval for this study was granted by the College of Medicine Research Ethics Committee, 195 Kamuzu University of Health Sciences (P.02/15/1677), the Liverpool School of Tropical Medicine 196 Research Ethics Committee (14.056) and the London School of hygiene and Tropical Medicine 197 (26839). Individual written informed consent, including consent for publication, was obtained from 198 each participant prior to study recruitment. 199 200

201 Descriptive analysis 202 A total of 2,067 ALWHIV aged 18-40y were enrolled in the study between 29 June 2015 and 9 203 August 2019. Among adults with non-missing data, 1,427 (69.0%, n=2,067) were females, 413 204 (23.3%, n=1,770) and 674 (38.1%, n=1,770) were from low and middle SES households, respectively, 205 1,156 (56.0%, n=2,066) were not living with a child <5y, 1,772 (98.5%, n=1,799) were on one of 206 Malawi’s first-line ART regimens, and 2,010 (97.2%, n=2,067) were using prophylactic 207 cotrimoxazole at recruitment. The median age was 33y (IQR: 28-37, n=2,067), median CD4+ count 208 was 252 cells/mm3 (IQR: 138-443, n=1,117), and median duration on ART at the time of study 209 recruitment was 3.0 years, (range: 0-17, n=1,530) (Fig. 1). 210 211 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 9 Using survey-aggregated data, serotype 3 comprised 77 (33.2%, n=232) of all VT serotypes 212 identified. Survey-aggregated data showed that carriage prevalence was 552 (26.7%, n=2,067) for 213 overall (VT+NVT) and 232 (11.2%, n=2,067) for VT. It also showed that overall and VT carriage 214 prevalence was 537 (37.7%, n=1,427) and 155 (10.9%, n=1,427) among females, 247 (38.6%, n=640) 215 and 77 (12.0%, = 640) among males, 195 (47.2%, n=413) and 59 (14.3%, n=413) in low SES, 276 216 (40.9%, n=674) and 85 (12.6%, n=674) in middle SES and 243 (35.6%, n=683), 73 (10.7%, =683) in 217 high SES households, 361 (39.7%, n=910) and 111 (12.2%, n=910) in adults living with a child <5y, 218 423 (36.6%, n=1,156) and 121 (10.5%, n=1,156) in adults living without a child <5y, 682 (38.5%, 219 n=1,778) and 199 (11.2%, n=1,778) in adults on a first-line ART regimen, 9 (33.3%, n=27) and 3 220 (11.1%, n=27) in adults on second-line ART regimen, 767 (38.1%, n=2,010) and 228 (11.3%, 221 n=2,010) in adults taking cotrimoxazole, 17 (29.8%, n=57) and 4 (7.0%, n=57) in adults not taking 222 cotrimoxazole (Fig 1). 223 224 225 Figure 1. Demographics and clinical characteristics of participants using aggregated data across eight 226 surveys. (A) Frequency of each VT in carriage; insert shows frequency of VT, NVT and no carriage. 227 (B) Number of adults in each annual age per survey with circle size proportional to total sample size. 228 The number of adults with VT, NVT, and no carriage living with (C) varying number of children <5 229 years or (D) across survey years. Notched box plots by serotype group representing participant 230 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 10 distribution by (E) age, (F) duration on ART and (G) CD4+ count. Proportion of serotype group by 231 (H) sex, (I) social economic status, (J) ART regimen and (K) Cotrimoxazole use. 232 233 Age- and time-dependent carriage prevalence estimates 234 Our GAM predicted a significant reduction in overall and VT carriage prevalence with increasing age 235 and time. Among older age categories, overall carriage prevalence was lower than the reference 236 younger adults aged 18-24y, with greatest reduction in adults aged 30-34y (-22.8%, 95%CI -34.1, -237 10.4). Likewise, VT carriage prevalence was lower in older than younger adults, with highest 238 reduction in adults aged 30-34y (-45.1%, 95%CI -61.8, -25.6). Across time, we estimated lower 239 overall (-38.2%, 95%CI -51.7, -23.6) and VT (-60.6%, 95%CI -79.1, -39.2) carriage prevalence in 240 2019 compared to 2015. In a sub-analysis, serotype 3 made up 22.6-34.7% (across age groups) and 241 18.9-38.2% (across time) of VT carriage prevalence (Table 1; Fig. 2). 242 243 Table 1. Age- and time-dependent overall and VT carriage prevalence, and relative difference in fitted carriage prevalence amon g ALWHIV on ART, 2015-2019 in Blantyre, Malawi. Observed overall carriage n/N (%) Modelled‡ overall carriage prevalence (95%CI) Relative difference† in overall carriage (95%CI) Observed VT carriage n/N (%) Modelled‡ VT carriage prevalence (95%CI) Relative difference† in VT carriage (95%CI) Age (years) 18-24 144/310 (46.5) 49.2 (40.9, 58.0) Reference 50/310 (16.1) 19.5 (13.0, 28.6) Reference 25-29 120/337 (35.6) 40.0 (33.8, 46.8) -18.7 -32.3, -4.1) 32/337 (9.5) 12.4 (8.7, 17.6) -36.4 (-58.0, -11.4) 30-34 213/585 (36.4) 38.0 (32.1, 44.5) -22.8 (-34.1, -10.4) 63/585 (10.8) 10.7 (7.4, 15.1) -45.1 (-61.8, -25.6) 35-40 307/835 (36.8) 38.0 (31.8, 45.0) -22.8 (-33.3, -11.2) 87/835 (10.4) 11.8 (8.1, 17.1) -39.4 (-55.9, -19.9) Year 2015 114/265 (43.0) 45.0 (37.4, 53.3) Reference 40/265 (15.1) 17.0 (11.4, 24.8) Reference 2016 218/494 (44.1) 44.7 (38.0, 52.0) -0.7 (-16.3, 16.8) 67/494 (13.6) 15.3 (10.6, 21.8) -10.0 (-37.7, 24.4) 2017 226/561 (40.3) 41.6 (35.2, 48.6) -7.6 (-22.1, 8.7) 69/561 (12.3) 13.1 (9.1, 18.8) -22.9 (-46.9, 6.8) 2018 156/450 (34.7) 34.8 (29.0, 41.3) -22.7 (-36.2, -7.7) 41/450 (9.1) 8.8 (5.9, 12.9) -48.2 (-67.3, -25.3) 2019 70/297 (23.6) 27.8 (22.0, 34.6) -38.2 (-51.7, -23.6) 15/297 (5.1) 6.7 (4.2, 10.6) -60.6 (-79.1, -39.2) ‡ Carriage prevalence was modelled by fitting a GAM to individual carriage trajectories adjusting for risk factors as described in Methods † Relative difference was computed by subtracting a GAM carriage prevalence estimate from the reference category and then divid ing the difference by the

category and then multiplying by 100%. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 11 VT - PCV13 vaccine serotype. ALWHIV - adults living with human immunodeficiency virus. ART – Antiretroviral therapy 95%CI - 95% confidence interval. Values in bold are statistically significant ( p<0.05). 244 245 Figure 2. Observed and fitted pneumococcal carriage prevalence curves using data from rolling, 246 prospective cross-sectional surveys in Blantyre, Malawi 2015-19. Number of carriage samples per age 247 group between 18 to 40 years (y) and survey time from 2015 to 2019 represented by open circles 248 radius proportion to total sample size with corresponding confidence intervals (vertical lines). P-spline 249 GAM fitted lines and confidence intervals (ribbons) for the (A) age- and (B) time-dependent carriage 250 prevalence stratified by overall carriage, vaccine serotypes (VT) carriage and VT carriage without 251 serotype 3 (VT-st3). 252 253 Factors associated with overall carriage prevalence 254 Overall carriage prevalence was only independently associated with SES, with adults in low SES 255 having 22% higher overall carriage than those in high SES (21.9, 95%CI 1.6, 43.7). In a sub-analysis 256 with age and time-stratification, our model predicted that being a younger (18-24y) adult in low SES 257 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 12 or living with a child aged <5y was significantly associated with higher overall carriage prevalence. 258 Significant associations with low SES and shorter ART duration were also seen with overall carriage . 259 Overall carriage prevalence in younger adults was significantly higher by 42% for those in low vs 260 high SES (41.8, 95%CI 12.5, 74.0) and 27% for those living with vs those living without a child <5y 261 (27.2, 95%CI 0.4, 57.4). Temporally, overall carriage prevalence was persistently higher by 50% in 262 2018 (49.5, 95%CI 13.4, 89.8) and 84% in 2019 (83.6, 95%CI 20.5, 167.4) in adults in the low vs 263 high SES, and higher by 35% in 2015 (35.4, 95%CI 0.9, 77.2) and 131% in 2019 (130.8, 95%CI 43.2, 264 255.4) in adults with shorter vs longer duration on ART (Table 2, Fig. 3, S1 Table). 265 266 Factors associated with VT carriage prevalence 267 Sex, SES, ART duration and living with a child <5y were not significantly associated with VT 268 carriage prevalence. However, with age- and time-stratification, our model of VT carriage outcome 269 predicted that being a younger (18-24y) or older (35-40y) adult living without a child <5y or being 270 older male significantly increased VT carriage prevalence. Temporally, living without a child <5y 271 remained a significant predictor of higher carriage prevalence in 2019. Living without vs with a child 272 <5y significantly increased VT carriage prevalence by 67% in younger adults 67.1 (95%CI 10.7, 273 140.5) and 41.0% in older adults (95%CI 1.5, 88.6). VT carriage prevalence was significantly higher 274 in older males than females (50.0, 95%CI 7.2, 100.7; Table 2, Fig. 3, S1 Table). 275 276 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 13 Table 2. Risk factors for age- and time-dependent overall and VT with serotype 3 carriage prevalence, and the relative differen ces in the fitted carriage prevalence between the reference group and comparative groups among ALWHIV on ART, 2015- 2019 in Blantyre, Malawi. Risk factors Overall % relative difference† (95%CI) % Relative difference† in age-dependent carriage prevalence (95%CI) % Relative difference† in time-dependent carriage prevalence (95%CI) 18-24y 25-29y 30-34y 35-40y 2015 2016 2017 2018 2019 Overall carriage Female vs male -4.8 (-20.9, 13.5) -18.1 (-35.7, 1.7) 16.3 (-15.9, 57.2) 19.0 (-8.5, 52.9) -10.3 (-25.5, 6.8) -13.5 (-35.5, 12.7) -6.1 (-23.9, 14.2) 7.1 (-14.5, 32.5) -12.9 (-33.8, 12.2) 16.8 (-27.4, 79.3) Low vs High SES 21.9 (1.6, 43.7) 41.8 (12.5, 74.0) 2.1 (-26.0, 32.9) 22.2 (-2.8, 49.2) 19.9 (-1.8, 42.9) 16.5 (-16.2, 52.2) 19.3 (-5.4, 45.8) 26.7 (-0.1, 55.2) 49.5 (13.4, 89.8) 83.6 (20.5, 167.4) ART <3y vs ART ≥ 3y 11.5 (-6.7, 31.5) 1.9 (-20.8, 28.4) 4.7 (-23.9, 40.6) 4.4 (-15.7, 26.8) 19.9 (-0.4, 41.8) 35.4 (0.9, 77.2) 3.9 (-15.7, 25.3) -0.7 (-19.4, 19.9) 26.2 (-4.1, 61.4) 130.8 (43.2, 255.4) With vs without child <5y 9.7 (-0.8, 29.1) 27.2 (0.4, 57.4) 2.6 (-22.9, 31.8) 0.3 (-19.5, 22.1) 8.9 (-9.2, 28.5) 8.7 (-18.4, 39.5) 1.1 (-17.9, 22.2) 7.4 (-12.4, 29.4) 13.8 (-14.3, 44.7) -7.6 (-41.8, 31.8) VT carriage Female vs male -18.6 (-64,9, 53.7) -16.6 (-51.2, 28.6) 26.8 (-42.8, 146.6) -5.1 (-47.8, 59.1) -33.3 (-55.6, -5.8) -21.4 (-58.3, 31.3) -18.0 (-48.9, 22.5) -13.3 (-47.6, 33.1) 1.1 (-54.4, 92.7) 40.3 (-72.9, 315.6) Low vs High SES 6.0 (-61.0, 91.2) -26.5 (-67.6, 19.6) 12.5 (-46.6, 85.4) 23.8 (-32.4, 90.8) 35.8 (-14.5, 93.2) 17.7 (-47.0, 94.6) -3.2 (-48.7, 47.9) 17.6 (-38.2, 80.9) 37.3 (-39.2, 133.2) 53.1 (-48.0, 229.1) ART <3y vs ART ≥ 3y 0.0 (-55.1, 77.4) -14.5 (-49.6, 32.3) -15.3 (-63.7, 65.3) 33.7 (-19.3, 102.1) -2.9 (-36.4, 35.0) 32.7 (-26.2, 114.5) -6.9 (-42.9, 36.2) -21.9 (-51.5, 13.6) 1.0 (-49.2, 67.6) 54.4 (-53.7, 255.6) With vs without child <5y 33.1 (-40.9, 142.5) -40.2 (-64.7, -9.9) -6.6 (-52.3, 54.3) 8.6 (-36.6, 67.6) -29.1 (-52.8, -0.5) 7.7 (-43.3, 73.9) 26.9 (-21.2, 89.1) 50.4 (-6.0, 125.0) 4.5 (-55.1, 78.6) -84.8 (-108.2, -57.9) † Relative difference was computed by subtracting a GAM carriage prevalence estimate of the reference category from the compara tor category and then dividing the absolute difference by the reference category then multiplied by 100%. ART: Antiretroviral therapy, ALWHIV: adults living with human immunodeficiency virus, CI: confidence intervals, y: year, SES: S ocial Economic Status score based a possession index which is calculated as a sum of positive responses for household ownership of each of the fifteen different functioning items such as watch, radio, bank account, iron (charcoal), sewing machin e (electric), mobile phone, CD player, fan (electric), bednet, mattress, bed, bicycle, motorcycle, car, and television. Middle and high SES were combined and named as high SES. 277 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 14 278 Figure 3. P-spline generalised additive model: Observed and fitted pneumococcal carriage prevalence 279 curves for each potential risk factor category using data from rolling, prospective cross-sectional 280 surveys in Blantyre, Malawi 2015-19. Nasopharyngeal samples across age groups from 18 to 40 years 281 (y) and time represented by open circles. Circle radius is proportional to total sample size with 282 corresponding confidence intervals (vertical lines). The coloured lines show P-spline GAM fitted 283 model and confidence intervals for age- and time-dependent carriage prevalence for overall carriage 284 (plots from first row) and vaccine serotypes (VT) carriage (plots from second row) stratified by risk 285 factor categories. 286 287 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 15

288 We used generalised additive models to estimate age- and time-dependent overall (VT+NVT) and VT 289 pneumococcal carriage prevalence and related risk factors in ALWHIV on ART. We analysed overall 290 and VT carriage separately to take into account the effect of a high uptake infant PCV13 programme 291 42. Overall and VT carriage declined with increasing age group and time, with VT carriage having a 292 faster decline (faster still if serotype 3 was excluded from VT). Our models predicted higher overall 293 carriage prevalence in younger adults from low SES and living with a child <5y, as well as those with 294 shorter duration on ART. Conversely, VT carriage prevalence was predominantly high in older males, 295 and younger and older adults not living with a child <5y. These findings suggest that the decline in 296 VT carriage prevalence across time in ALWHIV on ART is in part due to VT indirect protection from 297 vaccinated younger children, although it is imperfect in males or adults not living with younger 298 children, who may potentially have different routes of pneumococcal acquisition outside the 299 household. Reduction in overall carriage suggests a combined effect of infant PCV13 vaccination and 300 widespread use of antibiotics (i.e. cotrimoxazole) among adults in this setting. 301 302 Our analysis of risk factors for persistent carriage in ALWHIV on ART in the post PCV13 303

era in urban Blantyre extends our previous observations that focussed mainly on high 304 residual VT carriage and its determinants in PCV13-vaccinated and unvaccinated children 15,16. We 305 now show substantially high overall and VT carriage prevalence in ALWHIV during the earlier (45% 306 and 17%) than later (28% and 7%) years post infant-PCV13 introduction. VT carriage declined faster 307 than overall carriage, suggesting cumulative vaccine-induced community-level indirect protection 308 from infant PCV vaccination 43–45. The temporal reduction in VT carriage prevalence was even more 309 marked when serotype 3 was excluded (and included as NVT), supporting accumulating evidence of 310 the reduced effectiveness of PCV13 against serotype 3 36,37. 311 312 Higher overall and VT carriage prevalences among younger than older adults reported in this study 313 may suggest distinct high carriage acquisition risk in younger adults, partly supported by recent 314 evidence of higher rates of skin-to-skin contacts between younger adults and with other age groups in 315 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 16 urban Blantyre 46. The shorter median duration on ART amongst the younger adults as shown in S6 316 Fig may contribute to this residual pneumococcal carriage through incomplete immune reconstitution 317 at both the systemic and mucosal level 27,47,48. 318 319 Low SES neighbourhoods in urban Blantyre predominantly comprise high-density informal 320 settlements, relatively larger households and low rates of formal employment 28. Thus, substantial 321 overall carriage prevalence in younger adults from low SES suggests that factors associated with low 322 SES such as poorly ventilated and overcrowded houses with intense social contacts are reservoirs for 323 pneumococcal carriage in the PCV13 era 16,28,46,49. On the contrary, non-differential VT carriage 324 prevalence by household SES underlines an important role PCV vaccination plays to outweigh 325 infection risks in poor settings. We uncover a phenomenon where adults living with children 90%) infant PCV13 vaccination coverage 15, have 327 substantially lower VT but higher overall carriage prevalence suggesting some non-vaccine serotype 328 (NVT) replacement in adults within households, in line with evidence from rural Malawi and South 329 Africa 50,51. 330 331 In this setting, VT carriage prevalence was higher in older male than female adults. VT carriage 332 acquisition between mothers and their infants has been demonstrated previously in Malawi and South 333 Africa prior to infant-PCV introduction 20,21. Thus, our finding aligns with recent evidence in the same 334 setting showing strong intergenerational social mixing patterns between females and their potentially 335 PCV13-vaccinated younger children likely through parental or guardian roles 46. This suggests that in 336 the infant-PCV13 era, interruption of VT carriage transmission likely favors females than males. 337 338 Overall and VT carriage prevalence in ALWHIV on ART are heterogenous by age such that 339 epidemiological models for carriage that incorporate ALWHIV should stratify for age for precise 340 estimations. Our findings have policy implications in sub-Saharan African populations affected by 341 HIV as persistent VT carriage in ALWHIV may imply continued risk of VT-IPD 14. The indirect 342 impact on VT carriage of alternative infant-PCV13 vaccine strategies, including 2 primary doses with 343 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 17 a booster dose or double booster doses (i.e. 2+1 or 2+1+1), currently being tested to improve the 344 control of childhood disease, should also be further evaluated in ALWHIV 18. Indeed the 2+1 345 schedule, as implemented in South Africa, has generated indirect protection against IPD in 346 unvaccinated older children and ALWHIV 44,52. However, simply improving control of carriage in 347 young children to indirectly protect vunerable immunocompromised adults may be insufficient, 348 particularly in the context of a high local force of infection and a rapid waning of vaccine-induced 349 immunity 16,53. Furthermore, we provide evidence of heterogeneity in VT carriage prevalence with 350 males or adults living without <5y child in their homes being at highest risk of VT carriage in the 351 PCV era. Together, these data add weight to our viewpoint that as with many people living in high-352 income countries, targeted-pneumococcal vaccination should be considered in ALWHIV in LMICs. 353 354 We used a robust dataset with adequate samples to compute estimates for the overall, VT and risk 355 factor-dependent carriage prevalence. Nonetheless, there were some limitations to our work, including 356 limited data on risk factors such as viral load, use of tobacco, presence of other chronic co-357 morbidities, adherence to ART and history of antibiotics, which may independently influence carriage 358 dynamics 54. In addition, latex agglutination method used in the main analysis for single serotype 359 detection could underestimate our current prevalence estimates as compared to a more sensitive 360 microarray method for multiple serotype detection as show in S2 Fig. Finally, the systematic 361 recruitment of ALWHIV may be prone to bias if a cyclical pattern (unnoticeable here) is present in 362 the important characteristics of the individuals as they attend the ART clinic 55. 363 364 In conclusion, despite temporal reductions in overall pneumococcal carriage, the risk of VT carriage 365 and therefore subsequent pneumococcal disease remains high in ALWHIV. Efficient infant PCV 366 schedules that enhance indirect protection together with targeted-vaccination for ALWHIV should be 367 considered, along with other public health measures to further reduce VT carriage and disease. These 368 measures should be supported by robust surveillance to assess effectiveness and identify early 369 evidence of vaccine escape. 370 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 18

371 We thank the individuals who participated in this study and the local authorities for their support. We 372 are grateful to the study field teams (supported by Farouck Bonomali and Roseline Nyirenda). We are 373 grateful to the hospitality of the QECH ART Clinic, led by Ken Malisita. Our thanks also extend to 374 the MLW laboratory management team (led by Brigitte Denis) and the MLW data management team 375 (led by the late Clemens Masesa whose contribution we wish to particularly acknowledge). DT, KCJ, 376 JO, SF NF, RSH, and TDS are supported by the National Institute for Health and Care Research 377 (NIHR) Global Health Research Unit on Mucosal Pathogens and RSH is a NIHR Senior Investigator. 378 The views expressed in this publication are those of the authors and not necessarily those of the NIHR 379 or the Department of Health and Social Care. The MLW Programme is supported by a Strategic 380 Award from the Wellcome, UK 381 Author contributions 382 Conceptualization; DT, SF, NF, TDS, RSH 383 Data curation; DT, TDS, TM 384 Formal analysis; DT, SF 385 Funding acquisition; NF, RSH, TDS 386 Investigation; DT, TM, AK, JM, CB, TDS 387 Methodology; DT, SF 388 Project administration; TM, AK, JM, CB, CM, NF, RSH, TDS 389 Resources; NF, RSH, TDS 390 Software; DT 391 Supervision; SF, NF, KCJ, TDS 392 Validation; DT, TM, KCJ, AK, JM, CB, CM, JO, SF, NF, RSH, TDS 393 Visualization; DT 394 Writing – original draft; DT 395 Writing - review & editing; DT, TM, KCJ, AK, JM, CB, CM, JO, SF, NF, RSH, TDS 396 All authors read and approved the final manuscript. 397 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 19 398 Data availability 399 An R script that was used to analyse the datasets is available in the GitHub repository 41. 400 401 Competing interests 402 The authors declare no competing interests. 403 404 Role of the funding source 405 A project grant jointly funded by the UK Medical Research Council (MRC) and the UK Department 406 for International Development (DFID) under the MRC/DFID Concordat agreement, also as part of the 407 EDCTP2 programme supported by the European Union (Grant MR/N023129/1); and a recruitment 408 award from the Wellcome (Grant 106846/Z/15/Z). The MLW Programme is supported by a Strategic 409 Award from the Wellcome, UK. The National Institute for Health and Care Research (NIHR) Global 410 Health Research Unit on Mucosal Pathogens is supported using UK aid from the UK Government 411 (Grant 16/136/46). SF is also supported by a Sir Henry Dale Fellowship jointly funded by the 412 Wellcome Trust and the Royal Society (Grant 208812/Z/17/Z). The views expressed in this 413 publication are those of the author(s) and not necessarily those of the NIHR or the Department of 414 Health and Social Care. The funders had no role in study design, collection, analysis, data 415 interpretation, writing of the report or in the decision to submit the paper for publication. The 416 corresponding author and senior authors had full access to the study data, and together, had final 417 responsibility for the decision to submit for publication. 418 419 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 20

420 1. WHO. WHO Pneumococcus vaccines position paper . 85–104 421 https://www.who.int/immunization/policy/position_papers/pneumococcus/en/ (2019). 422 2. Simell, B. et al. The fundamental link between pneumococcal carriage and disease. Expert 423 Review of Vaccines 11, 841–855 (2012). 424 3. Bar-Zeev, N., Mtunthama, N., Gordon, S. B., Mwafulirwa, G. & French, N. Minimum Incidence 425 of Adult Invasive Pneumococcal Disease in Blantyre, Malawi an Urban African Setting: A 426 Hospital Based Prospective Cohort Study. PLOS ONE 10, e0128738 (2015). 427 4. Meiring, S. et al. HIV Infection and the Epidemiology of Invasive Pneumococcal Disease (IPD) 428 in South African Adults and Older Children Prior to the Introduction of a Pneumococcal 429 Conjugate Vaccine (PCV). PLOS ONE 11, e0149104 (2016). 430 5. Corcoran, M. et al. The epidemiology of invasive pneumococcal disease in older adults in the 431 post-PCV era. Has there been a herd effect? Epidemiology & Infection 145, 2390–2399 (2017). 432 6. Dwyer-Lindgren, L. et al. Mapping HIV prevalence in sub-Saharan Africa between 2000 and 433 2017. Nature 570, 189 (2019). 434 7. Frank, T. D. et al. Global, regional, and national incidence, prevalence, and mortality of HIV, 435 1980–2017, and forecasts to 2030, for 195 countries and territories: a systematic analysis for the 436 Global Burden of Diseases, Injuries, and Risk Factors Study 2017. The Lancet HIV 6, e831–e859 437 (2019). 438 8. Choko, A. T. et al. Uptake, Accuracy, Safety, and Linkage into Care over Two Years of 439 Promoting Annual Self-Testing for HIV in Blantyre, Malawi: A Community-Based Prospective 440 Study. PLOS Medicine 12, e1001873 (2015). 441 9. van Aalst, M. et al. Incidence of invasive pneumococcal disease in immunocompromised 442 patients: A systematic review and meta-analysis. Travel Medicine and Infectious Disease 24, 89–443 100 (2018). 444 10. Harries, A., Makombe, S., Libamba, E. & Schouten, E. Why Did the Scale-up of HIV Treatment 445 Work?: A Case Example From Malawi. Jaids Journal of Acquired Immune Deficiency Syndromes 446 57, (2 011). 447 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 21 11. Harries, A. D. et al. Act local, think global: how the Malawi experience of scaling up 448 antiretroviral treatment has informed global policy. BMC Public Health 16, 938 (2016). 449 12. Jahn, A. et al. Scaling-up antiretroviral therapy in Malawi. Bulletin of the World Health 450 Organization 94, 772–776 (2016). 451 13. VACFA. Immunization Schedules - Africa | Vaccines for Africa. 452 http://www.vacfa.uct.ac.za/immunization-schedules-africa (2018). 453 14. Bar-Zeev, N. et al. Impact and effectiveness of 13-valent pneumococcal conjugate vaccine on 454 population incidence of vaccine and non-vaccine serotype invasive pneumococcal disease in 455 Blantyre, Malawi, 2006–18: prospective observational time-series and case-control studies. The 456 Lancet Global Health 9, e989–e998 (2021). 457 15. Swarthout, T. D. et al. High residual carriage of vaccine-serotype Streptococcus pneumoniae after 458

of pneumococcal conjugate vaccine in Malawi. Nature Communications 11, 2222 459 (2020). 460 16. Lourenço, J. et al. Determinants of high residual post-PCV13 pneumococcal vaccine-type 461 carriage in Blantyre, Malawi: a modelling study. BMC Medicine 17, (2019). 462 17. Klugman, K. P. & Rodgers, G. L. Population versus individual protection by pneumococcal 463 conjugate vaccination. The Lancet 393, 2102–2104 (2019). 464 18. Thindwa, D. et al. Vaccine strategies to reduce the burden of pneumococcal disease in HIV-465 infected adults in Africa. Expert Review of Vaccines 0, 1–8 (2020). 466 19. WHO | Pneumococcal conjugate 3rd dose (PCV3) immunization coverage. WHO 467 http://www.who.int/gho/immunization/pneumococcal/en/ (2019). 468 20. Shiri, T. et al. Dynamics of Pneumococcal Transmission in Vaccine-Naïve Children and Their 469 HIV-infected or HIV-uninfected Mothers During the First 2 Years of Life. Am J Epidemiol 178, 470 1629–1637 (2013). 471 21. Heinsbroek, E. et al. Pneumococcal Acquisition Among Infants Exposed to HIV in Rural 472 M alawi: A Longitudinal Household Study. Am J Epidemiol 183, 70–78 (2016). 473 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 22 22. le Polain de Waroux, O. et al. Characteristics of human encounters and social mixing patterns 474 relevant to infectious diseases spread by close contact: a survey in Southwest Uganda. BMC 475 Infectious Diseases 18, 172 (2018). 476 23. le Polain de Waroux, O. et al. Identifying human encounters that shape the transmission of 477 Streptococcus pneumoniae and other acute respiratory infections. Epidemics 25, 72–79 (2018). 478 24. Neal, E. F. G. et al. Factors associated with pneumococcal carriage and density in children and 479 adults in Fiji, using four cross-sectional surveys. PLOS ONE 15, e0231041 (2020). 480 25. Neal, E. F. G. et al. Associations between ethnicity, social contact, and pneumococcal carriage 481 three years post-PCV10 in Fiji. Vaccine 38, 202–211 (2020). 482 26. Heinsbroek, E. et al. Persisting high prevalence of pneumococcal carriage among HIV-infected 483 adults receiving antiretroviral therapy in Malawi. Aids 29, 1837–1844 (2015). 484 27. Glennie, S. J. et al. Defective Pneumococcal-Specific Th1 Responses in HIV-Infected Adults 485 Precedes a Loss of Control of Pneumococcal Colonization. Clin Infect Dis 56, 291–299 (2013). 486 28. National Statistical Office Malawi. Malawi Population and Housing Census 2018 . 1–311 487 http://www.nsomalawi.mw/index.php%3Foption%3Dcom_content%26view%3Darticle%26id%3488 D226:2018-malawi-population-and-housing-489 census%26catid%E2%80%89%3D%E2%80%898:reports%26Itemid%E2%80%89%3D%E2%80490 %896 (2019). 491 29. Ministry of Health. Malawi Guidelines for Clinical Management of HIV in Children and Adults . 492 1–128 493 https://differentiatedservicedelivery.org/Portals/0/adam/Content/yb4xSSLvE0SW98_z7wTm_w/494 File/Malawi%20Clinical%20HIV%20Guidelines%202018%20(1).pdf (2018). 495 30. Satzke, C. et al. Standard method for detecting upper respiratory carriage of Streptococcus 496 pneumoniae: Updated recommendations from the World Health Organization Pneumococcal 497 Carriage Working Group. Vaccine 32, 165–179 (2013). 498 31. Newton, R., Hinds, J. & Wernisch, L. Empirical Bayesian models for analysing molecular 499 serotyping microarrays. B MC Bioinformatics 12, 88 (2011). 500 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 23 32. Swarthout, T. D. et al. Evaluation of Pneumococcal Serotyping of Nasopharyngeal-Carriage 501 Isolates by Latex Agglutination, Whole-Genome Sequencing (PneumoCaT), and DNA 502 Microarray in a High-Pneumococcal-Carriage-Prevalence Population in Malawi. Journal of 503 Clinical Microbiology 59, (2020). 504 33. Kamng’ona, A. W. et al. High multiple carriage and emergence of Streptococcus pneumoniae 505 vaccine serotype variants in Malawian children. BMC Infectious Diseases 15, 234 (2015). 506 34. Stekhoven, D. J. & Bühlmann, P. MissForest—non-parametric missing value imputation for 507 mixed-type data. Bioinformatics 28, 112–118 (2012). 508 35. Eilers, P. H. C. & Marx, B. D. Flexible Smoothing with B-splines and Penalties. Statistical 509 Science 11, 89–102 (1996). 510 36. Linley, E., Bell, A., Gritzfeld, J. F. & Borrow, R. Should Pneumococcal Serotype 3 Be Included 511 in Serotype-Specific Immunoassays? Vaccines 7, 4 (2019). 512 37. Usuf, E. et al. Persistence of Nasopharyngeal Pneumococcal Vaccine Serotypes and Increase of 513 Nonvaccine Serotypes Among Vaccinated Infants and Their Mothers 5 Years After Introduction 514 of Pneumococcal Conjugate Vaccine 13 in The Gambia. Clin Infect Dis 68, 1512–1521 (2019). 515 38. Hens, N. et al. Modeling Infectious Disease Parameters Based on Serological and Social Contact 516 Data: A Modern Statistical Perspective. (Springer Science & Business Media, 2012). 517 39. Kohavi, R., Longbotham, R., Sommerfield, D. & Henne, R. M. Controlled experiments on the 518 web: survey and practical guide. Data Min Knowl Disc 18, 140–181 (2009). 519 40. R Core Team (2018). R: A language and environment for statistical computing. R Foundation for 520 Statistical Computing, Vienna, Austria. https://www.r-project.org/. 521 41. Thindwa, D. R code and data for Age- and time-dependent risk factors for pneumococcal carriage 522 in HIV-positive adults on ART in the infant PCV era in Blantyre, Malawi. 523 https://github.com/deusthindwa/Pneumo.carriage.adults.hiv.malawi (2021). 524 42. Mvula, H. et al. Predictors of Uptake and Timeliness of Newly Introduced Pneumococcal and 525 Rotavirus Vaccines, and of Measles Vaccine in Rural Malawi: A Population Cohort Study. PLOS 526 ONE 11, e0154997 (2016). 527 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 24 43. Nzenze, S. A. M. C. et al. Temporal Changes in Pneumococcal Colonization in a Rural African 528 Community With High HIV Prevalence Following Routine Infant Pneumococcal Immunization. 529 Journal 32, 1270–1278 (2013). 530 44. Nzenze, S. A. et al. Imputing the Direct and Indirect Effectiveness of Childhood Pneumococcal 531 Conjugate Vaccine Against Invasive Pneumococcal Disease by Surveying Temporal Changes in 532 Nasopharyngeal Pneumococcal Colonization. Am J Epidemiol 186, 435–444 (2017). 533 45. Roca, A. et al. Effects of Community-Wide Vaccination with PCV-7 on Pneumococcal 534 Nasopharyngeal Carriage in The Gambia: A Cluster-Randomized Trial. PLOS Medicine 8, 535 e1001107 (2011). 536 46. Thindwa, D. et al. Social mixing patterns relevant to infectious diseases spread by close contact 537 in urban Blantyre, Malawi. 2021.12.16.21267959 538 https://www.medrxiv.org/content/10.1101/2021.12.16.21267959v1 (2021) 539 doi:10.1101/2021.12.16.21267959. 540 47. Glennie, S. J. et al. Impaired CD4 T cell memory response to Streptococcus pneumoniae precedes 541 CD4 T cell depletion in HIV-infected Malawian adults. PLoS ONE 6, e25610 (2011). 542 48. Zhang, L. et al. Humoral immune responses to Streptococcus pneumoniae in the setting of HIV-1 543 infection. Vaccine 33, 4430–4436 (2015). 544 49. Dherani, M., Heinsbroek, E., Tafatatha, T., Chartier, R. & Bruce, N. Household Air Pollution and 545 Pneumococcal Carriage in 6 Months Old Children in Malawi – MSCAPE Study. ISEE 546 Conference Abstracts (2018) doi:10.1289/isee.2017.2017-558. 547 50. Heinsbroek, E. et al. Pneumococcal carriage in households in Karonga District, Malawi, before 548 and after introduction of 13-valent pneumococcal conjugate vaccination. Vaccine 36, 7369–7376 549 (2018). 550 51. Cohen, C. et al. Effectiveness of the 13-valent pneumococcal conjugate vaccine against invasive 551 pneumococcal disease in South African children: a case-control study. The Lancet Global Health 552 5, e359–e369 (2017). 553 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint 25 52. Madhi, S. A. & Nunes, M. C. The potential impact of pneumococcal conjugate vaccine in Africa: 554 Considerations and early lessons learned from the South African experience. Human Vaccines & 555 Immunotherapeutics 12, 314–325 (2016). 556 53. Flasche, S., Lipsitch, M., Ojal, J. & Pinsent, A. Estimating the contribution of different age strata 557 to vaccine serotype pneumococcal transmission in the pre vaccine era: a modelling study. BMC 558 Medicine 18, (2020). 559 54. Thindwa, D. et al. Estimating the contribution of HIV-infected adults to household pneumococcal 560 transmission in South Africa, 2016–2018: A hidden Markov modelling study. PLOS 561 Computational Biology 17, e1009680 (2021). 562 55. PHAST. Methods of sampling from a population. Health Knowledge 563 https://www.healthknowledge.org.uk/public-health-textbook/research-methods/1a-564 epidemiology/methods-of-sampling-population (2010). 565 566 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 16, 2022. ; https://doi.org/10.1101/2022.05.12.22274986doi: medRxiv preprint