Study area
The Ethiopian sector of the East African Rift system extends for more than 1000 km in a north east-south west to north- south direction from the Afar depression, at Red Sea-Gulf of Aden junction, southwards to the Turkana depression. It is a long and narrow strip of low-lying plain land in between the highlands. It stretches from the north eastern part of the country to the southern border with Kenya, and divides the highland masses into two, the central and eastern highlands of Ethiopia (4). Groundwater is an important source of drinking water and the dominant source for domestic supply in the region where surface waters are scarce and seasonal (17-19). This area of the country usually associated with high level of fluoride in groundwater because of volcanic and basaltic rocks, which are more likely to release a high concentration of fluoride (10, 17, 20).
Data Sources and Search Strategy
Peer reviewed journal articles which have reported the mean concentration of fluoride in ground water source and prevalence of dental fluorosis in Ethiopian Rift valley were searched via scientific databases, libraries and the journals themselves. MEDLINE/PubMed, Cochrane library and Google scholar databases were the main sources for the studies. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guideline (21) was used to screen articles (Fig 1). The following searching terms with Boolean operators were used to search PubMed; fluoride OR f-OR fluorosis OR drinking AND water OR ground AND water AND rift valley AND Ethiopia. In addition, we tried to check references of searched papers to obtain additional articles which is not identified in the databases.
Inclusion and Exclusion Criteria
This review included all studies on the general population and measured outcomes indicative of fluoride concentration (dental fluorosis) in ground water in Ethiopian rift valley. Studies reported fluoride concentration (the minimum, maximum, mean and standard deviation in mg/l) were considered. In addition, the daily water consumption and fluoride intake per body weight also extracted from some studies. Regarding dental fluorosis, articles were selected based on fluorosis report with different severity level (mild, moderate and sever). Studies published in full text or abstract and published only in the English language (to avoid any mistake in the translation process) were included. Peer-reviewed and undertook laboratory works and the reported measurements of laboratory tests also considered during screening (Table 1). All articles which did not meet the proposed criterion were excluded.
Table 1: Eligibility criteria
Inclusion criteria
|
Exclusion criteria
|
- Country and setting: Ethiopian rift valley
|
- The article did not contain original data or observations
|
- Study design: Cross-sectional
|
- Study subjects were not humans
|
- Measurement: The concentration of fluoride in ground water and prevalence of dental fluorosis
|
- Fluoride concentration was not measured or estimated
|
- Types of article: Peer-reviewed and published in English language
|
- Dental fluorosis was not reported using the three-severity level
|
- Publication year: up to December, 2018
|
- The mean and standard deviation of Fluoride concentration was not reported
|
- Undertook laboratory works: Reported type of laboratory tests and the respective result
|
|
- Measurable outcomes: dental fluorosis (Dean’s index)
|
- Being none related to fluoride in ground water supplies and dental fluorosis
|
Quality assessment
The Newcastle-Ottawa Scale (NOS) which is adapted for cross-sectional studies (additional file 1) was used to assess the quality of included studies. Four investigators (HD, AB, ZA and AM) assessed the quality of included studies based on the efforts made by the authors of each primary studies to minimize the risk of error and bias. Inconsistencies between the quality assessor were discussed to reach consensus. But in most case the risk of bias for the prevalence of fluorosis across these studies was unjustifiable. To minimize the chance of error in data extraction, a pre-determined and standardized data extraction form was prepared and piloted with all review team members prior to the actual data extraction.
Data extraction
Data extraction of all included papers was conducted by the three authors independently (HD, ZA and AM). These data extractors have taken practical training on the process of sourcing and recording relevant details from the primary studies included in the systematic review. Tailored Microsoft excel was used to ensure consistent data extraction, whilst reducing bias and improving validity. The following variables were extracted from studies: author, publication year, study population, water sample, sample size of human participants, age group and mean age of the participants, prevalence of dental fluorosis (mild, moderate and sever) and the concentrating of fluoride (minimum, maximum, mean and standard deviation in mg/l). But we didn’t contact study authors to obtain data needed for the analysis that were not reported in the published articles. The detail characteristics of included studies are shown in (Table 2) and Table 3).
Data analysis
All relevant studies which provided data on fluoride concentration in ground water and crude prevalence of dental fluorosis or numbers of cases and study participants were included in the meta-analysis. Some studies reported the numerical concentration without calculating the mean fluorine concentration. In this case, the mean concentration of fluoride was calculated by summing up all concentration and divided by number of water samples. Whereas the prevalence of dental fluorosis for individual studies was determined by multiplying the ratio of dental fluorosis cases to sample size by 100.The estimation of pooled fluoride concentration and prevalence of dental fluorosis was done using CMA 2.0 and MetaXL version 5.3 software. With the assumption that true effect sizes exist between eligible studies, the random effects model was used to determine the pooled prevalence, mean fluoride and 95% CIs. A summary (pooled) estimate is calculated as a weighted average from selected studies as follows;
weighted average = sum of (estimate X weight)
sum of weights
Heterogeneity was evaluated using the Cochran’s Q test and I2 statistics. Significant heterogeneity was declared at I2 > 50% and Q-test (P < 0.10).
Publication bias and sensitivity analysis
The recurrence analysis under different assumptions to examine the impact of these assumptions on the results was done. Funnel plots were drawn to assess the possibility of publication bias and it indicated potential for publication bias. We plotted the studies’ logit event rate and the standard error to detect asymmetry in the distribution.
Heterogeneity
The included eleven studies were assessed for heterogeneity. The differences between studies in the characteristics of their populations, in their study designs and quality and the variation of effects between studies was checked using common sense, graphical and statistical tests. Forest plot was drawn and the overlapping confidence intervals was checked.