Oral Health Status of Children and Adolescents Living with HIV Undergoing Antiretroviral Therapy: A Systematic Review and Meta-Analysis

Antiretroviral therapy (ART) increases the survival of HIV-infected children, but might also bring in oral health-related side effects and increase their risks of oral diseases. The review compared the oral health status of children living with HIV (CLWH) undergoing ART with healthy controls. Dual independent screening and study selection from four electronic databases and manual searches, data extraction, risk of bias assessment, and quality-of-evidence evaluation with Grading of Recommendations Assessment Development and Evaluation were performed. Twelve studies were included in qualitative and quantitative analysis. CLWH taking ART had a significantly higher prevalence of periodontal diseases (OR = 3.11, 95% CI 1.62–5.97), mucosal hyperpigmentation (OR = 20.35, 95% CI 3.86–107.39), and orofacial-related opportunistic infections than healthy controls. No significant differences regarding caries prevalence and tooth development were identified. Those with CD4+ T-cell counts below 250 cells/mm3 were more likely to manifest opportunistic infections, while medication duration had minimal influence on the prevalence of orofacial opportunistic infections. The current findings did not identify HIV and antiretroviral status as predisposing factors to dental caries, but affirmed the associated increased risk of periodontal diseases, mucosal hyperpigmentation and candidiasis.


Introduction
With the implementation of HIV prevention campaigns and advancement in antiretroviral therapy (ART), the new incidence of children living with HIV (CLWH) had declined significantly [1]. Over 70% of HIV-infected women could have access to effective ART during pregnancy, delivery, and breastfeeding [1]. The risk of vertical transmission of HIV has been reduced to less than 6% [1].
For CLWH, the current treatment approach involves the prescriptions of combined ART or so-called highly active antiretroviral therapy (HAART) [2]. HAART refers to a mixed prescription of ART agents, for instance, two nucleoside reverse transcriptase inhibitors (NRTI) with either a protease inhibitor (PI) or non-nucleoside reverse transcriptase inhibitor (NNRTI) [2]. With the early onset of HAART, a majority of CLWH can live an asymptomatic life and continue to thrive [3].
However, ART may bring along adverse effects to the surviving CLWH, impacting their general health and quality of life. The most commonly reported side effects include skin rashes, anemia, hepatotoxicity, and other detrimental damages to the gastrointestinal, metabolic, and renal systems [4,5]. Oral health-related side effects associated with ART have also been reported. Reduced salivary flow among patients taking protease inhibitors and didanosine [6] might predispose them to higher risks of dental caries and other oral diseases. Susceptibility to dental caries may be exacerbated due to prolonged use of sweetened liquid oral medication among pediatric patients [7]. Oral ulcerations have been identified among

Materials and Methods
This systematic review was conducted following the PRISMA guidelines and was registered on PROSPERO (registration number CRD42019148245). The following PECO(S) statements were proposed: (P)-Participants were children and adolescents below 18; (E)-Exposure was those undergoing ART and HAART treatment; in this present review, ART refers to all the treatment regimens used to treat HIV, while HAART specifically refers to a combination of three or more antiretroviral medications [13].
Oral hygiene and periodontal health status 3.
Dental development (S)-Studies included were case-control observational studies with full text reports available in English.

Search Strategies
Utilizing MeSH terms and broad keywords (Appendix A), a strategic literature search was performed systematically. Four electronic databases (Ovid Embase, Ovid MEDLINE, Pubmed, and Scopus) were searched from inception to 29 July 2022. Reference lists of past relevant literature reviews were also screened to identify any additional pertinent reports.

Study Selection
Two reviewers (PPYL and NZ) independently assessed the eligibility of retrieved articles based on their titles and abstracts. Any disagreement was resolved by consensus or consulting the third reviewer (HMW). Cohen's kappa coefficient (κ) was calculated to determine the agreement between reviewers.

RoB
The studies by Ponnam et al. (2012) and Subramaniam and Kumar (2015) were considered as of severe risk of bias due to confounding and overall RoB as no matching of socioeconomic status of controls were performed. The other ten studies were assessed as of low risk of overall RoB, Figure 2.

RoB
The studies by Ponnam et al. (2012) and Subramaniam and Kumar (2015) were considered as of severe risk of bias due to confounding and overall RoB as no matching of socioeconomic status of controls were performed. The other ten studies were assessed as of low risk of overall RoB, Figure 2.

Dental Caries Experience
Two studies with 1028 case-control subjects contributed to the outcomes regarding dental caries prevalence (DMFT/dmft/DMFS/dmfs > 0). Ponnam Figure 3. Rajonson et al. (2017) reported that CLWH also had a significantly higher median DMFT/dmft/DMFS/dmfs than HIV-uninfected children. Using a zero-inflated negative binomial regression model to evaluate potential confounders, HIV status was found to be associated with both higher caries prevalence and median DMFT/dmft/DMFS/dmfs among children below 12 years old; whereas only significantly higher caries prevalence but not median DMFT/dmft/DMFS/dmfs was found for children above 12 years.

Oral Hygiene and Periodontal Status
For oral hygiene status, only one eligible study [31] compared and reported the simplified oral hygiene index of CLWH with healthy controls, where no significant differences were found between the two groups.
For periodontal diseases, Bosco and Birman (2002) and Ponnam et al. (2012), with 250 case-control subjects, compared the prevalence of periodontal diseases between CLWH and healthy controls. Both studies reported significantly higher prevalence of periodontal diseases among CLWH and adolescents with ART. However, both studies did not specify the indices they used in determining gingivitis and periodontitis. Bosco and Birman (2002) further grouped the CLWH based on their CD4+ T-cell counts and degree of suppression and found gingivitis only prevailed in children with moderate to severe suppression (below 999 cells/mm 3 ). Despite consistent findings, moderate heterogeneity is shown (OR = 3.11, 95% CI 1.62-5.97, p = 0.001; I 2 = 69.7%, p = 0.069) in Table 2 and

Oral Hygiene and Periodontal Status
For oral hygiene status, only one eligible study [31] compared and reported the simplified oral hygiene index of CLWH with healthy controls, where no significant differences were found between the two groups.
For periodontal diseases, Bosco and Birman (2002) and Ponnam et al. (2012), with 250 case-control subjects, compared the prevalence of periodontal diseases between CLWH and healthy controls. Both studies reported significantly higher prevalence of periodontal diseases among CLWH and adolescents with ART. However, both studies did not specify the indices they used in determining gingivitis and periodontitis. Bosco and Birman (2002) further grouped the CLWH based on their CD4+ T-cell counts and degree of suppression and found gingivitis only prevailed in children with moderate to severe suppression (below 999 cells/mm 3 ). Despite consistent findings, moderate heterogeneity is shown (OR = 3.11, 95% CI 1.62-5.97, p = 0.001; I 2 = 69.7%, p = 0.069) in Table 2 and Figure 4.

Saliva Immunoglobulins Quantity
One study compared and reported the salivary IgA (SIgA) concentration [30]. There was no significant difference in the total SIgA among HIV-ART, HIV-HAART children, and healthy controls, but there was a significantly higher concentration of anti-C albicans SIgA in the former two groups.

Dental Development
Three studies [26,28,29] used either Nolla's method or Willems' method to assess dental age. The age of the included subjects in the three studies ranged from 4-16 years old, with even distributions of males and females. Fernandes et al. (2007) and de  reported no significant difference in dental development between the two groups. Holderbaum et al. (2005) reported that the mean dental age assessed by Nolla's method of HIV-infected males had no significant difference with the chronological age, while healthy male controls had their chronological age significantly overestimated. Whereas the dental age of HIV-infected females obtained with Nolla's method was significantly lower than their chronological age, however, the same underestimation of chronological age was noticed in the control group.

Other Associated Factors
Meta-regression could not be conducted due to the lack of raw data. Other confounding factors including CD4+ T-cell counts and duration of the medications were hence evaluated separately. Baghirath (2013) and Divakar (2015) reported that CLWH having CD4+ T-cell counts below 250 cells/mm 3 were nearly two times as likely to have HIV-related oral lesions (OR = 1.916, 95% CI 0.613-2.513, p = 0.001; OR = 1.999, 95% CI 0.129-3.688, p = 0.001 respectively).
Two studies [22,27] evaluated and compared the effect of the duration of HAART on the prevalence of opportunistic infections. The opportunistic infections investigated include linear gingival erythema, hairy leukoplakia, angular cheilitis, and oral ulcers. Except for candidiasis (OR = 17.93, 95% CI 3.69-83.12, p = 0.002); no significant differences in the prevalence of other opportunistic infection manifestations were found between those taking ART below three years and those taking it three years or more. However, fewer children who had taken ART not less than three years had CD4+ T-cell counts below 250 cells/mm 3 (Table 2).

Discussion
CLWH had been reported to have poorer oral health status [33]. However, conclusions were mostly drawn from cross-sectional studies without head-to-head comparisons with their healthy counterparts. Most cross-sectional studies were conducted in developing countries, the increased caries prevalence and severity might be contributed by other confounders; for instance, socioeconomic status and water fluoridation.
Divergent results were found when comparing the caries experience between HIV-ART/HAART and healthy individuals, with one study showing higher caries prevalence among HIV-ART/HAART individuals [9] while the other found no significant difference [31]. The effect estimate was provided by 1-2 small-scale studies with moderate to high RoB. Due to serious risks of bias, inconsistencies, or imprecisions, the certainties of evidence were severely compromised, precluding valid conclusions to be drawn in Supplementary Material Table S1. The current evidence does not suggest CLWH under medical management are more prone to dental caries and the disease itself is no predisposing factor to caries. Other risk factors, for instance, inadequate oral hygiene and cariogenic diet, might play a more fundamental role in the disease progression. More high-quality case-control studies are warranted to determine if CLWH are more susceptible to dental caries.
Consistent results demonstrated that HIV-ART/HAART individuals were more likely to have mucosal hyperpigmentation. ART was associated with elevated production of á-melanocyte-stimulating hormone(á-MSH), which might be responsible for the increased melanin production and coincided with the current findings [34]. Nonetheless, the treatment duration seems to have minimal influence on hyperpigmentation based on the pooled data of two small-scale studies available. Other than being drug-induced, oral hyperpigmentation was also found in other CLWH without taking any medications. These idiopathic lesions appeared as brown-black macules with well-defined margins, and melanin was only identified in keratinocytes of the basal cell layer or in extracellular foci [34]. One proposed theory was that the melanocytes were stimulated during HIV infection, resulting in immunopathologic changes in the oral mucosa [35], but further validations of such theories remain scarce. Further research is warranted due to the low certainty of the evidence.
Inconclusive results were also found on whether ART treatment would affect dental mineralization [26,28,29]. Dental maturity or dental age has been a debatable method to assess chronological age, as dental development markedly varies across different ethnic groups. Additionally, inaccuracies and overestimation of chronological age by Nolla and Willems method had been reported in the literature, ranging from 0.05 to 1.15 years and 0.06 to 0.26 years respectively [36,37]. The reported errors between dental age and chronological might not be because of the medications, but the errors of the assessment method itself.
It is not surprising to find that opportunistic infections were more prevalent among HIV-infected individuals when compared with healthy controls. The most common and distinctive opportunistic infections which were only found among HIV-infected individuals were oral candidiasis. Other opportunistic infections, such as linear gingival erythema, oral hairy leukoplakia, and angular cheilitis were not common in both healthy and HIV-infected individuals. The occurrence of oral candidiasis was more likely to be associated with the drop in CD4+ T-cell counts. Moreover, children with the presence of oral candidiasis were five times more likely to experience dental caries than those without [38], which might be contributed to the capability of C. albicans to acclimate and thrive in a wide range of pH [39], especially in acidic environment created by the cariogenic biofilm. Frequent screening and detection of oral candidiasis might be necessary to assess the disease status of HIV and the treatment progress.
This review was conducted following PRISMA guidelines [40]. The strengths include the use of both qualitative and quantitative synthesis, sensitivity analyses, and comprehensive evaluation of quality evidence using the GRADE approach [18].
The limitations include the inevitable exclusion of non-English translated reports, exclusion of subjects over 25, or non-extractable data. Due to small sample sizes and limited number of relevant studies, the difficulty in detecting heterogeneity increased. Although a higher p-value of 0.1 instead of the usual 0.05 was used for each meta-analysis, it might still not be sensitive enough given the clinical, methodological, and statistical diversity of the included studies [19]. Meanwhile, publication bias and meta-regression could not be conducted due to limited studies and the inability to pull relevant data from reports respectively. For future research, more high-quality case-control studies with matched controls controlling for sociodemographic and other confounding factors are warranted to provide information about the oral health status of CLWH undergoing ART.

Conclusions
The present review identified a significantly higher prevalence of periodontal diseases, hyperpigmentation, oral candidiasis, and ulcerative stomatitis among CLWH undergoing ART compared with healthy controls. However, no significant difference regarding caries prevalence and severity, oral hygiene status, saliva immunoglobulins, and dental development was found between CLWH and adolescents undergoing ART with healthy controls. The underlying factors for opportunistic infections might be attributed to the drop of CD4+ T-cell counts below 250 cells/mm 3 . However, due to the dearth of high-quality, case-control studies, more definitive statements regarding the oral health status of CLWH undergoing ART cannot be established.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijerph191912864/s1, Table S1: GRADE summary of findings table for the primary outcome (caries prevention and arrest) and secondary outcome.
Author Contributions: P.P.Y.L., N.Z., H.M.W. and C.K.Y.Y. were involved in study conception and design. P.P.Y.L. and N.Z. were responsible for data extraction, qualitative synthesis, and assessment of studies. Quantitative synthesis and result analyses were performed by P.P.Y.L., whom also drafted and wrote the manuscript, in consultation with N.Z., C.K.Y.Y. and H.M.W. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available in the supplementary materials.