Paediatrica Indonesiana p-ISSN 0030-9311.
e-ISSN 2338-476X.
Vol.
No.
DOI: https://doi.
org/10.
14238/pi62.
Original Article Virological failure of first-line antiretroviral therapy in children living with HIV in Indonesia and associated factors Nia Kurniati1,2.
Zakiudin Munasir1.
Pramita Gayatri1.
Evy Yunihastuti3.
Budiman Bela4.
Anggraini Alam5 Abstract Background The World Health Organization (WHO) recommends viral load (VL) monitoring for HIV patients on antiretroviral therapy (ART).
However, availability of VL monitoring in lowincome countries remains limited.
Objective To investigate factors associated with virological failure in HIV-infected children treated without routine VL monitoring.
Methods This cohort study was done in children living with HIV (CLHIV) registered at Cipto Mangunkusumo General Hospital from 2004 to 2021.
Viral load monitoring was not routinely done.
Subjects with at least one VL result after 6 months on ART were included in the study.
Virological failure was defined as a VL of >1,000 copies.
SubjectsAo data were obtained from medical records, laboratory reports, and dispensing pharmacies.
Statistical analysis was done following survival analysis with hazard ratio.
Results There were 384 children who had at least 1 VL result after ART was initiated.
Median age at diagnosis was 30 months.
Length of follow-up ranged from 6 to 216 months, with a mean frequency of VL monitoring of 0.
7 times/person/year.
Most subjects were already in clinical stages 3 and 4 .
8%).
75% met severe immunodeficiency criteria.
Virological failure was found in 45.
8% of subjects after a median of 33 months on first-line ART, yielding an incidence of 3.
3 per 1,000 person months.
Independent associated factors were age at diagnosis of <60 months (HR 1.
95%CI
13 to 2.
, severe immunodeficiency (HR 1.
95%CI 1.
15 to .
, referral cases (HR 1.
95%CI 1.
23 to 2.
, and WHO clinical staging 3 (HR 1.
95%CI 0.
995 to 3.
and 4 (HR 95%CI 1.
034 to 4.
Subjects with virological failure had lower weight-for-age z-scores .
interquartile range (IQR) -3.
003 to -0.
and height-for-age z-scores .
IQR -2.
902 to -1.
at the time of failure.
Conclusions In HIV-infected children treated without routine VL monitoring, age at diagnosis <60 months, severe immunodeficiency.
WHO clinical stage 3 and 4, and referral from other centers were associated with virological failure.
[Paediatr Indones.
62:295-303.
DOI: https://doi.
org/10.
14238/pi62.
295-303 ].
Keywords: HIV.
virological failure.
uman immunodeficiency virus (HIV) infection is a major public health problem In 2020, there were 37.
million people living with HIV (PLHIV), 5 million new HIV infections and 680,000 deaths from AIDS-related causes.
1 An estimated 7 million children worldwide .
ge <15 year.
were living with HIV and more than 80% of them were living in sub-Saharan Africa, of whom only an estimated 49% received antiretrovirals.
The use of antiretroviral therapy (ART) that combines two or more drugs has significantly reduced morbidity and AIDS-related mortality among infected 4,5 The goal of ART is to suppress viral replication, resulting in decreased HIV-RNA VL.
The World Health Organization (WHO) recommends routine viral load monitoring for all PLHIV on ART to assess treatment response and disease progression.
However, availability of VL monitoring remains From the Department of Child Health1 and Doctoral Program in Medical Sciences2.
Department of Internal Medicine3, and Department of Microbiology 4,Universitas Indonesia Medical School/Dr.
Cipto Mangunkusumo General Hospital.
Jakarta, and Department of Child Health.
Universitas Padjadjaran Medical School/Dr.
Hasan Sadikin Hospital.
Bandung.
West Java5.
Indonesia.
Corresponding author: Nia Kurniati.
Dr.
Cipto Mangunkusumo Hospital.
Jl.
Diponegoro 71.
Jakarta.
Indonesia.
Email: niawidj@gmail.
Submitted June 11, 2022.
Accepted October 26, 2022.
Paediatr Indones.
Vol.
No.
September 2022 A 295 Nia Kurniati et al.
: Virological failure of first-line ART in children living with HIV in Indonesia and associated factors scarce, particularly in low- and middle-income countries due to limited laboratory facilities and trained personnel.
9 In Indonesia.
VL monitoring is not routinely supported.
Missed opportunities to do routine viral monitoring lead to delays in identifying treatment failure and switching to second-line ART combinations.
Virological failure and the development of drug resistance have become major issues in children living with HIV (CLHIV) due to weight-based dosing, poor tolerability of drugs, and suboptimal adherence.
10,11
Studies in sub-Saharan Africa showed that the prevalence of virological failure was significantly higher among children compared to adults and was reported to be 32.
1% in Ugandan, 37% in Kenyan, and 66% in Malawian children based on cohort and cross-sectional studies of 12-24 months of therapy.
In South Africa, the incidence rate of virological failure was 18.
7 per 100 person years follow up, while India, as AsiaAos representative, reported a 29% prevalence of viral failure.
9,15 Several factors that are known to increase the risk of virological failure on CLHIV include younger age at diagnosis or treatment
WHO stage 3 and 4, lower CD4 count, nevirapine (NVP)-based regimen, and treatment
10,11,16-21
This study aimed to determine the incidence of treatment failure with first-line ART in situations of limited VL monitoring in Indonesia, as well as to identify the predictors of treatment failure in order to assist clinicians in monitoring their patients more Methods This retrospective cohort study included patients from the Dr.
Cipto Mangunkusumo General Hospital registry observed from 2004 to 2021.
Dr.
Cipto Mangunkusumo General Hospital is the top national referral hospital in the capital region of Jakarta.
Indonesia, with an HIV service established in the early 2000s.
Universal ART coverage started in 2004 and any PLHIV who met the indication criteria was The ART regimens were given according to the Indonesian Guideline of HIV Management, adapted from WHO guidelines.
The first national guideline was published in 2008, with subsequent 296 A Paediatr Indones.
Vol.
No.
September 2022 revised versions published in 2013, 2013, and 2019, incorporating continuous adjustments according to the global guidelines.
6 Drug doses were based on the childAos body weight.
Children were reviewed every month after ART initiation and every 2-3 months for those whose HIV infections were under control.
Children aged <18 years living with HIV/AIDS who had their first-line ART initiation between January 2004 and March 2021 and had at least one viral load measurement after six months of ART initiation were included.
The diagnosis of HIV infection was confirmed by virological test using polymerase chain reaction (PCR) of deoxyribonucleic acid (DNA) or ribonucleic acid (RNA) HIV for patients aged <18 months and serological test for patients aged Ou18 months (Cobas TaqMan.
Roche.
Basel and GenXpert.
Cepheid.
Sunnyvale.
Californi.
Children who transferred from other hospitals and initiated ART with complete information available were included in the study.
CLHIV with incomplete data were excluded.
The outcome variable was virological failure of CLHIV after starting ART.
The sociodemographic characteristics collected were age at diagnosis, sex, the childAos primary caregiver, disclosure of HIV status to the CLHIV prior to virological failure, and antiretroviral (ARV) treatment history .
hether nayve or referra.
Baseline clinical factors were WHO clinical stage and baseline CD4 according to age Treatment-related factors were type of initial regimen, form of medication use, time to treatment failure, age at treatment failure, timing of switch to second-line treatment after failure, type of secondline regimen, adherence.
HIV status disclosure, and nutritional status at treatment failure.
Time to failure was measured in months from the point of ART initiation until the point of first-line treatment failure.
We determined adherence before and after failure in cross-sectional fashion.
Virologic failure was diagnosed as VL >1,000 copies/ml after at least 6 months of ART treatment.
Baseline data were taken at the first clinic visit and included age at diagnosis, sex, the childAos primary caregiver.
WHO clinical stage at diagnosis, and whether the patient was a nayve or referral case.
Adherence to highly active antiretroviral therapy (HAART) was measured for those with viral load >1,000 copies/ml, measured subjectively, and Nia Kurniati et al.
: Virological failure of first-line ART in children living with HIV in Indonesia and associated factors expressed on a scale of 0 to 100 according to a visual analog scale (VAS).
A VAS of 91-100 was considered as good, 81-90 as moderate, and <80 as poor adherence.
CD4 count was categorized according to the WHO age-appropriate classification to describe immunosuppression level: children aged <1 year with a CD4 count of <1,500 cells/mm3, 1-5 years with a CD4 count of <750 cells/mm3, and >5 years with a CD4 count of <350 cells/mm3 were categorized as First-line ART regimens consisted of three The backbone was either nevirapine (NVP), efavirenz (EFV), lopinavir/ritonavir (LPV/.
or other drugs.
The other two drugs were made up of a combination of either zidovudine (AZT), stavudine (D4T), didanosine (DDI), tenofovir (TDF), lamivudine .
TC), or emtricitabine (FTC) and abacavir (ABC).
The forms of medications used are specific to Indonesia.
Since most ART are available as adult tablets, we categorized the medication forms as crushed tablet, scored tablet, or full form tablet.
When pediatric fixed drug combination was available and used, patients recorded it as full form tablet.
Time to treatment failure was calculated from the date of first ART dispensing from the pharmacy and the date of virological failure.
Nutritional status, comprising body weight, body height and body mass index (BMI), was expressed as age- and sex-specific z-scores according to the WHO Child Growth Standards.
23 Z-scores were determined using WHO AnthroA software (WHO.
Genev.
for children aged <5 years and WHO AnthroPlusA (WHO.
Genev.
for children aged 5-18 years.
The measurements collected were those done within two weeks before or after the date of treatment failure.
Weight-forage z-score (WAZ), height-for-age z-score (HAZ), and BMI z-score (BAZ) of <-2 were considered as Data were collected and collated from the hospital electronic data registry, physical medical records, laboratory reports, and ARV dispensing lists from the hospital pharmacy.
Data were initially collected using Microsoft Excel 2021A (Microsoft.
Redmond.
Washingto.
Completeness of the data was checked manually.
Data coding and entry were done using SPSS version 28 (IBM.
Armonk.
New Yor.
Data consistency was repeatedly checked by the investigators.
Retrospective data collection encompassing a period of 18 years has a high possibility of information We expected data to be missing in several The minimum required sample size was determined using a single population proportion formula, with the following statistical assumptions:
an incidence rate of 1.
5 per person months based on a study by Makatini et al.
15 in South Africa, a margin of error of 5%, and a Za/2 of 1.
96, which was the corresponded to a 95% confidence interval .
%CI).
Precision was determined at 1% and time of observation (T) was 216 months, resulting in a minimum sample size of 267 subjects.
Tables were used to present descriptive results.
Additionally, frequencies, percentages, proportions, and median .
ith interquartile rang.
were used to summarize the study population characteristics.
Two standard statistical methods were employed: the Kaplan-Meier method .
non-parametric metho.
and Cox proportional hazards regression model .
semiparametric metho.
Stata v.
16 software (StataCorp.
College Station.
Texa.
was used to generate substitute values for missing data .
mputate missing dat.
to use in multivariate analysis.
The Kaplan-Meier method was used to estimate the rate of virological failure at different time points.
The incidence of virological failure was calculated in person-years of observation.
The Chi-square test was used to examine bivariate associations between independent variables and virological failure, using Stata and SPSS.
We further analyzed all bivariate associations with a P value of <0.
25 using Cox proportional hazards analysis to determine significant predictors of virological failure.
The results are presented as hazard ratio (HR) with 95%CI.
The level of significance was set at p<0.
This study was approved by the Health Research Ethics Committee.
Universitas Indonesia Medical School/Dr.
Cipto Mangunkusumo General Hospital.
To protect patient confidentiality, all records were anonymized and de-identified prior to analysis.
All procedures in this study were conducted in accordance with the Declaration of Helsinki, as revised in 2013.
Results There were 684 children with HIV registered from January 2003 to March 2021, which was the last Paediatr Indones.
Vol.
No.
September 2022 A 297 Nia Kurniati et al.
: Virological failure of first-line ART in children living with HIV in Indonesia and associated factors Eligible patients .
Excluded Follow up time <6 months .
Not on ART .
No viral load after receiving RT .
Enrolled subjects .
Figure 1.
Subject recruitment recruitment date for new patients in order to allow for 6 months of treatment observation.
The study observation ended on 31 December 2021.
excluded 178 children who had not been under our care for at least six months, either due to death or having only one or two consultation visits with less than six months of follow up.
Thirty-one children who were not on ART were excluded.
Furthermore, 91 children were excluded because, although they met the six-month ART criteria, they had no data on VL monitoring.
Hence, a total of 384 CLHIV were included in the final analysis (Figure .
Median age of HIV diagnosis was <3 years .
months for those with failure and 30 months for those without failur.
According to absolute CD4 count for age, most subjects were already in WHO clinical stage 3 or 4 with severe immunodeficiency .
%).
Twenty-one percent of new cases in the study were referred from other HIV care centers.
Most subjects had started ART one to 11 years before referral to our hospital, usually for failure diagnosis.
Almost all subjects were infected through a perinatal route, but only four had received ART for prevention during the perinatal period.
SubjectsAo mean frequency of VL examination was 0.
73 times per person per year.
A single parent or both parents were the primary caregivers for most subjects, with another 38% of subjects under the care of grandparents or other blood The basic sociodemographic characteristics of subjects before ART initiation are shown in Table 1.
The median time lag between diagnosis and ART initiation was 7 months in the failure group and 5 months in the non-failure group.
Since pediatric ART formulations were scarce, adult ARV tablets were usually crushed or scored.
Although 47.
2% of 298 A Paediatr Indones.
Vol.
No.
September 2022 subjects took crushed tablets .
owder for.
, they used the full tablet form at later ages.
Table 2 shows the characteristics of ART treatment in subjects.
The longest follow-up time was 216 months .
Disclosure was important for CLHIV once they reached the late school-age or teen years.
In our study almost half of subjects were still at an inappropriate age for disclosure (Table .
More than 75% of subjects used nevirapine (NVP) as first-line ART backbone, which was in line with Indonesian guidelines for those aged <3 years.
Use of efavirenz is contraindicated in this age group.
A small number of subjects used ritonavir-boosted lopinavir (LPV/.
as first-line backbone.
most had started ART overseas or had an allergy to nevirapine.
In the past 5 years, we changed the backbone to LPV/r whenever possible, as the appropriate, child-friendly form was available.
Overall, almost half .
8%) of the participants met our definition of virological failure .
iral load >1,000 copies/mL) after at least six months of ART consumption .
8%), as defined by the 2016 WHO Guideline, measured 2 times apart with adherence In 2021, the WHO guideline was modified to allow a one-time VL result >1,000 copies/mL to define virological failure, which was more feasible in this study.
6 Adherence measurements before and after the date of viral failure were available for 117 subjects, with 92% having good adherence (VAS 91-.
Data for adherence was missing for those in care before 2017, and this data was not imputed.
At virological failure, median weight-for-age z-score was -1.
92 (IQR
81 to -3.
, classified as undernutrition, and median height-for-age z-score was -2.
05 (IQR -1.
, classified as short stature.
Nia Kurniati et al.
: Virological failure of first-line ART in children living with HIV in Indonesia and associated factors Table 1.
Sociodemographic characteristics of subjects Variables Total .
Virological failure .
Non-failure .
Gender, n (%) Male Female 198 .
Source of transmission, n (%) Perinatal Blood products Sexual 380 .
Age at diagnosis Median, months IQR (Q3-Q.
Age at diagnosis by group, n (%) O60 months >60 months 312 .
WHO clinical stage before ART, n (%) Grade 3 and 4 Grade 1 and 2 299 .
Type of cases, n (%) Referral Nayve 81 .
Primary caregiver, n (%) Parents Blood relatives Other .
oster parent.
Absolute CD4 based on age Immunocompromised Not immunocompromised 289 .
Disclosure status prior to viral failure .
=269, missing data= .
Yes Partial No HIV disclosure due to age < 12 y.
Unknown Nutritional status at virological failure or last VL* .
=344, missing data .
Body weight (WAZ) Body height (HAZ) BMI 72 (-0.
578 to -2,.
73(-0.
775 to -2.
25 to -1.
92 (-0.
81 to -3.
05 (-1.
04 to -2.
04 to -1.
54 (-0.
33 to -2.
495 (-0.
50 to -2.
36 to -1.
*Median Z-score (Q3-Q.
Using a life table calculation, the incidence rate of virological failure was 3.
3 per 1,000 person months of follow-up.
Most subjects with virological failure were male .
3%).
In subjects with virological failure, median age at HIV diagnosis was 26 (IQR 50-.
Median overall time to failure was 33 months.
The proportion of failure was higher in referral cases .
/81.
1%).
Out of all failure subjects, 29.
5% were referral cases.
Diagnosis of failure was the reason of referral in most subjects.
Before performing analysis on the association of determinants with virological failure, we imputed missing data on body weight and body height in 40 out of 384 subjects .
4%) based on age at diagnosis and WHO clinical stage.
Both bivariate and multivariate analyses were used to determine factors associated with virological failure.
On bivariate analysis, age at diagnosis.
WHO clinical stage.
ART treatment history .
eferral or nayv.
, drug formulation, and absolute CD4 count were associated with virological failure with a P value of <0.
25, therefore becoming candidate variables for the final regression model (Table .
Multivariate Cox regression analysis revealed four statistically significant determinants for virological Paediatr Indones.
Vol.
No.
September 2022 A 299 Nia Kurniati et al.
: Virological failure of first-line ART in children living with HIV in Indonesia and associated factors Table 2.
Characteristics of ART in subjects Variables Total .
Virological failure .
Non-failure .
Age at ART initiation, months Median IQR (Q3-Q.
ART regimens, n (%) NVP-based EFV-based LPV/r-based Other 137 .
Form of ART drugs .
=380, missing data=4 Crushed tablet Full or scored tablet 137 .
Adherence, .
=117, missing data= .
* Poor Moderate Good *Measured only in failure subjects, thus % is out of 117 total subjects.
Adherence level: poor (VAS<.
, moderate (VAS 80-.
, good (VAS 91-.
Table 3.
Multivariate Cox regression analysis for virological failure among CLHIV Variables Multivariate analysis Bivariate analysis Initial model Final model HR .
%CI)
P value HR .
%CI) P value aHR .
%CI) Gender 03 .
76 to 1.
14 to 1.
P value Age at diagnosis <60 month 39 .
92 to 2.
058 to 2.
13 to 2.
Stage 4 338 .
165 to 4.
Stage 3 207 .
016 to 4.
938 to 3.
034 to 4.
964 to3.
995 to 3.
Stage 2 963 .
908 to 4.
Stage 1 790 .
823 to 3.
815 to 3.
Referral cases 74 .
26 to 2.
337 to 2.
336 to 2.
ARV formulation 28 .
89 to 1.
423 to 2.
Parents 81 .
33 to 1.
Blood relatives 11 .
45 to 2.
19 to 2.
341 to 2.
WHO clinical staging:
Primary caregiver Other .
oster parents, orphane.
Absolute CD4 based on age 662 .
192 to 2.
806 to1.
Nutritional status BMI Z-score <0.
failure in the final model.
Children who were aged <60 months at diagnosis were prone to treatment failure (HR 1.
95%CI 1.
13 to 2.
P=0.
Referral cases were more likely to have virological failure compared to nayve cases (HR 1.
95%CI
336 to 2.
P=0.
Baseline absolute CD4 count categorized as immunocompromised (HR 1.
95%CI 1.
341 to 2.
P=0.
and severe WHO clinical stage .
tage 3 and .
were also significant predictors for virological failure (Table .
300 A Paediatr Indones.
Vol.
No.
September 2022 Discussion This study provides important information about virological failure status among CLHIV in Jakarta.
Current practice on treatment failure monitoring in Indonesia focuses on clinical deterioration or immunological failure, which are considered late 24 Since 2016, the WHO has proposed annual viral load monitoring to detect earlier treatment Nia Kurniati et al.
: Virological failure of first-line ART in children living with HIV in Indonesia and associated factors Determination of ideal virological failure was a challenge for our cohort, who underwent less than 1 test per person per year.
Ideally, after one VL result of >1,000 copies/ml, the viral load test should be repeated, as it supports adherence.
However, program constraints and lack of funding lead to few or no follow-up testing.
Even a single viral load test result of >1,000 copies/ml may be considered as failure, especially in low-and middle-income countries with nucleoside/nucleotide reverse transcriptase inhibitors (NNRTI) as backbone first-line ARV.
25 The timing for follow-up VL tests is important, since there is evidence of emerging viral resistance in older regimens.
Hence, the recommendation of 6-month, 12-month, and yearly testing after initiating ARV, thereafter.
22 Use of integrase inhibitor (INSTI) may change the need for an exact time of repeated viral load testing, since INSTI is considered to have higher resistance barrier.
Prevalence of virological failure in our study .
8%) was higher than reported in three studies in Africa and one study in India.
12-15 This difference could have been due to follow-up periods and study The follow-up time for our cohort was 6 months to 18 years, which was longer than in other Our unique position as the referral center for pediatric HIV care in the greater Jakarta region resulted in a distinctive group of subjects, whether they were nayve or had prior ART experience.
The proportion of failure was higher in referral subjects .
%), which was not previously reported in any study.
This is consistent with the diagnosis of treatment failure being the main reason for referral to our center.
Such a diagnosis implies that subjects probably had immunological failure or clinical failure that occurred after virological failure and before they were referred to our hospital.
The incidence density of treatment failure in our study was 3.
3 children per 1,000 person-months, which was higher compared to that in a South African study reporting an incidence density of 18.
7 per 100 person-years .
quivalent to 1.
5 per 1,000 personmonth.
15 The mean time from first-line therapy to virological failure was 29 months in our study, similar to that of 30 months reported by an Ethiopian study.
For the last 15 years in Indonesia.
ARV backbone regimens have been limited to nevirapine and efavirenz.
both of which were reported to be ineffective among African populations with high perinatal exposure.
15 However, exposure risk was not a determinant of virological failure in our study, since almost all subjects had missed perinatal prevention.
In subjects treated with nevirapine, the proportion of failure was much higher compared to those treated with efavirenz, but still comparable to findings of other studies reporting virological failure.
18,19 However, univariate analysis of first-line backbone regimens showed no statistically significant differences in hazard Weight and height in CLHIV usually improve after ART.
In our study, nutrition data was extracted at failure diagnosis.
This approach differed from another study which used baseline data at diagnosis in order to measure its influence on failure occurrence.
study in South Africa showed that WAZ and HAZ at baseline had no relationship to virological failure.
15 In other studies, wasting and stunting were not always measured as separate factors, although WHO stage 3 or 4 was reported as an important factor in virological failure, with HIV wasting as one criteria.
10,14,15,17,20
Findings of WAZ <-1 and HAZ <-1.
9 showed that clinical failure had already taken place.
Ineffective regimen use and lack of scheduled VL monitoring may explain this late detection of virological failure.
Determinants associated with virological failure in the Cox proportional hazards (PH) multivariate final model were initial absolute CD4 counts categorized as immunocompromised, age at diagnosis <60 months, referral cases, and clinical stages 3 and 4 (Table .
The proportion of immunocompromised status at baseline in our study was comparable to that in other studies.
Form of medication .
rushed or scored vs.
whole table.
a HR of >1, but the p value did not agree with the confidence interval.
Since pediatric ART formulations were scarce, adult ARV tablets were usually crushed or scored.
The practice of scoring or crushing ARV tablets may have led to a sub-standard pharmacological effect.
These specific results could be pertinent to hospital characteristics which could be used to improve care during clinic Our study used a strong design and a longer follow-up duration to estimate the survival time and predictors of viral failure.
However, our limitations associated with incomplete data may have led to underestimating or overestimating virological failure.
Missing data on variables may have limited the range Paediatr Indones.
Vol.
No.
September 2022 A 301 Nia Kurniati et al.
: Virological failure of first-line ART in children living with HIV in Indonesia and associated factors of variables and the number of children that could be included in multivariate models.
Imputation was only done for nutritional data, which accounted for 10.
of missing data.
In conclusion, the prevalence and cumulative incidence of virological failure was high in our CLHIV Children who were <60 months of age at diagnosis, belonged to WHO clinical stage 3 and 4, had low absolute CD4 count, and were referral cases had a significantly higher risk of treatment failure.
This observational study showed that early HIV case finding remains inadequate, since the median age at diagnosis was around 3 years.
Late diagnosis may lead to treatment failure in the long term.
Based on our findings, areas for improvement in HIV care practices based on our findings are early diagnosis, more effective ART, routine VL monitoring, and timely switching to a second-line treatment regimen.
Conflict of interest None declared.
Acknowledgements The authors would like to thank Dr.
Melati Adi Prameswari and Dr.
Nida Ghitha for their help in gathering data from the registry and medical records.
Salfia Lastari for assistance in data management and statistical analysis, as well as Dr.
Andrian Wiraguna.
Dr.
Rizqi Amalia, and Dr.
Dina Muktiarti for their valuable input.
References