The systematic review protocol was registered within the “Open Science Framework” (OSF, DOI 10.17605/OSF.IO/DE7BH). This systematic review followed the “Preferred Reporting Items for Systematic Reviews and Meta-Analyses” (PRISMA) guidelines (20). The inclusion criteria were formulated based on the following PICOS tool questions (21): P (population) = healthy Muslim adults willing to fast DR; I (intervention/exposure) = exposure to RIF; C (Comparison): DR and outside Ramadan [i.e., before-Ramadan (BR) and after-Ramadan (AR)]; O (Outcome): SFR and saliva metabolic parameters; and S (Study design): all original articles written in English. No restrictions were applied in terms of study design, setting, country, or period. Publications not in compliance with the purpose of this systematic review as well as those not representing original research (i.e., reviews, editorials, qualitative papers, case reports, case series, and letters to editors) were not included.

An online literature search was performed using two databases: PubMed and Scopus from their inception up to 15 July 2021. For PubMed, the search was carried out using a strategy employing the combination of the following two “Medical Subject Headings” (MeSH) terms: Saliva AND Fasting. As for Scopus, the previous two terms were searched for in the article titles, abstracts, and/or keywords. In addition, the reference lists of the included articles were checked. All the authors involved in this review agreed on the articles to be included in this systematic review.

The process of articles selection is outlined in Figure 1. Duplicate articles were eliminated using End-Note X9 library. Titles of the remaining articles were independently appraised during the initial online literature search for studies by two of the authors (AB and MK in the authors’ list) to check for their relevance to the searched topics. Abstracts of these titles were then read to determine if the studies met the inclusion criteria. The studies whose abstracts met the inclusion criteria were then read in full-text format to determine their eligibility and therefore retention. Two authors (AB and MK in the authors’ list) conducted the study selection process for this review, with discrepancies being checked by a third author (HBS in the authors’ list), if necessary.

Data from the retained studies were extracted using a format including the population, the parameters being investigated, the periods during which the parameters were collected, and the significant findings. Data were extracted, reviewed, and analyzed by two authors (AB and MK in the authors’ list). Extracted data were then verified by a third author (HBS in the authors’ list). Discrepancies in data collection were resolved through discussion.

Methodological quality assessment was performed using the Joanna Briggs Institute (JBI) critical appraisal tool, precisely the checklist for cohort studies (https://joannabriggs.org/last visit: 4 March 2022). The checklist appraises the following areas: recruitment, exposure measurement, reliability of exposure measurement, confounding factors identified, strategies to deal with confounding factors, participants free of outcome at the onset of the study, validity and reliability of outcome measurement, follow-up timeframe reported, follow-up completion, strategies utilized to deal with incomplete follow-up, and appropriate statistical analysis. The checklist included the following 11 items: 1. Were the two groups similar and recruited from the same population? 2. Were the exposures measured similarly to assign people to both exposed and unexposed groups? 3. Was the exposure measured in a valid and reliable way? 4. Were the confounding factors identified? 5. Were the strategies to deal with confounding factors stated? 6. Were the groups/participants free of the outcome at the start of the study (or at the moment of exposure)? 7. Were the outcomes measured in a valid and reliable way? 8. Was the follow-up time reported and sufficient to be long enough for outcomes to occur? 9. Was follow-up complete, and if not, were the reasons for loss to follow-up described and explored? 10. Were the strategies to address incomplete follow-up utilized? and 11. Was the appropriate statistical analysis used? These items are scored as either yes, no, unclear, or not applicable. Two reviewers (AB and MK in the authors’ list) independently scored the retained studies, with discrepancies being resolved through discussion. If discrepancies could not be resolved through discussion, a third author (HBS in the authors’ list) intervened to reach consensus. The risk of bias in the studies was judged to be low (“yes” scores > 70%), moderate (50 ≤ “yes” scores between ≤ 69%), and high (“yes” scores < 49%) (22).

The search process yielded 675 articles, of which 202 were duplicated. Among the 473 remaining papers, 465 were excluded based on the title and the abstract. When screening the references lists of the remaining eight articles (23–30), one additional paper was added (31). After assessing full-text articles for eligibility, three articles were excluded (23–25). Consequently, six articles were retained (26–31). The search results are presented in Figure 1.

The retained six studies were assessed for methodological quality (Table 1). All the studies have a high score of bias (i.e., final score ranging from 9.1 to 36.4%). Items 2, 5, and 10 were rated as not applicable for all the studies. No study reported data regarding items 4, 7, and 11. Six (26–31), five (26–29, 31), four (27–29, 31), one (29), and one (31) studies included information regarding items 6, 8, 3, 9, and 1, respectively.

Table 2 exposes the main characteristics and methodological points of the retained studies. The latter were published between 2004 (26) and 2020 (29, 30). The studies were conducted in Saudi Arabia (26), Turkey (27), United Arab Emirates (29), and Iran (28, 30, 31). The study design was not reported in three studies (27–29). In the remaining studies, three designs were applied: observational design with repeated measures (26), case-control (31), and descriptive analytical research design (30). All the studies (26–31) opted for convenience samples. The Ramadan year was omitted in three studies (28, 30, 31). Only one study (29) mentioned the number of fasting days DR. Only one study (28) reported the average ambient temperature, which was around 15°C. Only three studies reported the mean fasting duration [i.e., 12 h (26), 15 h (29), 17 h (27)]. No study reported data with regard to the Ramadan season, the average ambient pressure, or the average ambient humidity.

The number of evaluation sessions was two (27, 29, 30), three (26, 28), and four (31). Five studies (26–30) opted for a session BR with different periods applied [i.e., 1 day BR (28), 1 week BR (27, 29, 30), 2 weeks BR (26)]. Only one study opted for a session AR (i.e., 7 days AR) (31). The number of sessions DR was one (27, 29, 30), two (26, 28), and four (31), and different periods were retained [i.e., 1 week (26), first 10 days (31), 10–20 second days (31), third week (26, 30), 21–29 last days (31), 25th day of Ramadan (27, 28), last day of Ramadan (28, 29)].

Two-hundred twenty-nine participants fasting DR were included. The sample sizes varied from 8 (26) to 75 (30). Three studies included mixed population of males and females (27, 28, 30), two studies included only males (28, 31), and the participants’ sex was not reported in one study (26). Four studies included healthy participants (26–28, 31), one study involved both overweight and obese participants (29), and one study omitted to report the health status of the included participants (30). The included participants were: students (31), employees in a factory (28), staff of a training and research hospital (27), and nurses (30). Several non-inclusion/exclusion criteria were applied. They were related to habits [e.g., smoking (26, 28), alcohol-use (26), addiction to caffeinated beverages (26)], medication-use (26, 27, 29), some health complaints [e.g., sleep complaints (26)], acute diseases [e.g., upper respiratory tract infections (27), severe infections (31)], chronic conditions [e.g., unhealthy teeth or mouth (28), oral diseases (28, 31), internal diseases (28), endocrine diseases (29), diabetes mellitus (29, 30), metabolic disorders (30), cardiovascular diseases (29, 30), hearing impairment (30), headache (30), psychiatric shock (30), unspecified (27)], previous surgeries [e.g., bariatric surgery (29), head surgery (30)], pregnancy (29), weight management program (29), and job experience <1 year (30). Only one study highlighted that no participant practiced fasting as routine and voluntary rituals before the month of Ramadan (29). In one study (29), participants were asked to continue their regular diet during non-fasting hours, and not to alter their habitual physical exercise levels BR or DR. Participants’ ages varied from 24.2 ± 2.3 (31) to 59 (27) years. Participants’ weight and body mass index were reported in two (27, 29) and three (26, 27, 29) studies, respectively.

Eight different saliva parameters were evaluated: SFR (31), glucose (31), melatonin (26), cortisol (29, 30), immunoglobulin A (IgA) (27), uric-acid (28), alkaline phosphatase (ALP) (28), and aspartate amino-transferase (AST) (28). The numbers of saliva sampling were one (27, 28), two (29, 30), and three (26, 31). Different times of saliva sampling were chosen. In some studies, fixed times were applied [e.g., mid-night (26), 6h00 (30), 8h00 (26), between 11h00 and 13h00 (29), 16h00 (26)]. In some other studies, a minimum of hours of fasting was needed [e.g., 6 (31), 8 (28), 12 (27)]. Four studies reported that they opted for unstimulated saliva (27–29, 31), and only two studies reported the duration of saliva collection [e.g., 2–5 (31) and 5 (27) min]. The volume of the collected saliva (in mL) was highlighted in four studies [e.g., 2 (30), 3 (28, 31), and 5 (26)]. One study omitted to report the equipment used to analyze the saliva outcomes (29). Before saliva collection, participants were asked to rinse their mouths with water (26–31) and to avoid: (i) coughing or throat clearing into the collection tube (26), (ii) consuming caffeine and substances containing melatonin or melatonin precursors (26), (iii) eating, drinking, and smoking (29, 30), (iv) brushing (30), and (v) using oral hygiene products (29).

Table 3 presents the main results of the six retained studies.

SFR decreased by 10% DR (31). DR, the lack of gustatory stimulation decreases the stimulation of salivary glands, therefore, SFR may decline. The autonomic nervous system controls SFR and the secretion of various salivary compounds (32). Stimulation of this system induces modifications in salivary secretions and SFR (33). In Ramadan, sedentary activity with minimal orofacial movement and metabolism slowing down in body tissues cells, including oral cavity cells, may explain the low stimulation of the autonomic nervous system (28). This hyposalivation can cause malodor, especially DR (34). Since saliva works to moisten the mouth, to neutralize acids produced by plaque, and to clean bacteria and food particles from the mouth, any salivary modifications create a suitable environment for aerobic and anaerobic bacteria that coat several sites in the oral cavity, notably the dorsum of the tongue (35). Overall, it has been shown that oral microflora modifications taking place DR may lead to malodor, even if other factors are involved (36).

Melatonin in saliva or plasma is an indicator of the timing of the circadian clock (37). According to Bahammam (26), the sleep hormone follows the same circadian rhythm both BR and DR. This means that melatonin secretion is low during the daytime, while the highest levels are released at night, but its level drops significantly from baseline (26). This variation may be due to the sleep habits modification DR (36). Nevertheless, this outcome should be considered with caution because of the small sample size in the study (n = 8) (26).

Cortisol is a hormone produced by the adrenal glands (38). Cortisol plays an essential role in balancing blood glucose and releasing sugar from the body’s stores in response to increased energy demands (39). Cortisol has an important role in the metabolism of fats and proteins as well as in the circadian rhythm regulation (38). This hormone is usually measured in the morning (7–9 a.m.) because it reaches a peak at this time (40). DR, external sources of glucose are reduced due to fasting. Consequently, salivary glucose concentration drops significantly (31). Thus, we can “speculate” that cortisol levels in saliva may rise to regulate glucose levels, however; the latter mechanism is not that straightforward and has to be elucidated by further research.

Dehagi et al. (30) reported that when participants are exposed to RIF and noise, which is another stress source, salivary cortisol levels increase. In addition to its glycemic effects, cortisol is also liberated during the stress periods in order to allow the body to adapt to an emotional or physical shock by mobilizing additional energy sources. The contradictory results of the studies of Al-Rawi et al. (29) and Dehagi et al. (30) may be due to methodological reasons, notably the study design and population, and the lack of information about the timing and duration of sleep in one study (30). It should be highlighted that many people in various Islamic countries may change their sleep rhythm during the Holy month. Indeed, their nighttime sleep duration is reduced compared to non-fasting days (41), in addition to the dietary patterns’ changes (42).

Alkaline phosphatase and aminotransferase are usually measured together to investigate the hepatic, cardiovascular, and renal functions (43). ALP is a protein produced by various cell types (e.g., polymorphonuclear leukocytes, osteoblasts, macrophages, and fibroblasts) within the alveolar bone and/or the salivary glands (44, 45). ALP can be a salivary biomarker of periodontal diseases and caries (46), as it interferes in the balance of the remineralization-demineralization cycle since it is primarily involved in calcium and phosphate binding (47). It seems that the function of ALP relatively depends on the salivary pH and buffering capacity (48). Khalighefar et al. (28) reported that ALP rebounds during the middle of Ramadan compared to BR. Although ALP increase may suggest much more susceptibility to dental caries and/or oral diseases, it is believed that this fluctuation is not so critical to lead to an illness. AST is an enzyme involved in the metabolism of several tissues and organs (49). Khaleghifar et al. (28) indicated that AST activity in fasting volunteers decreases significantly DR. This decrease can be related to the fact that fasting reduces the metabolism of body tissues cells, including oral cavity cells, thus leading to reduced SFR during fasting (28). Uric-acid is the ultimate product of the metabolic breakdown of purines, which are the nitrogenous bases in DNA and RNA (50). It is involved in healing and defense (50). Khaleghifar et al. (28) reported that uric-acid decreases DR since the metabolism is reduced (28). In contrast, several studies have shown that blood uric-acid increases during RIF (51–54). According to studies reported in the literature, despite the shifts in metabolic interactions among the organs producing uric-acid, AST or ALP, we cannot conclude on the effects of RIF on these enzymes because of the scarcity of these studies in addition to the limitations of the unique retained study investigating those parameters (28).

Salivary glucose DR plunges from baseline, especially in the first 10 days (31). First, this is expected because of food restriction for 4 weeks. Secondly, this fact is interesting and beneficial for oral health. Actually, both cariogenic bacteria and Candida use glucose for their development and survival (55, 56). This dysbiosis enhances the proliferation of these bacteria and dental biofilm development (46, 57). A recent study investigated the effect of different salivary glucose concentrations on dual-species biofilms of Candida albicans and Streptococcus mutans (58). The authors reported that higher salivary glucose increases counts of Candida albicans (58). It is possible that the higher levels of IgA detected in saliva BR can be attributed to the greater colonization of the oral cavity by Candida albicans due to the higher salivary glucose levels during that period compared to DR.

Salivary IgA has an important role in mucosal immunity. Its levels increase in case of oral mucosa infection, such as candidiasis. It allows inhibiting the adherence of candida to epithelial cells (59, 60). In contrast, the decrease in those salivary IgA levels does not necessarily suggest that the participant is more susceptible to oral infection onset, since a salivary IgA concentration threshold is absent (27). Subsequently, authors suggested that RIF results in neither severe immunological disturbances nor adverse impact on health (27). Some remarks related to the usefulness of salivary IgA in real practice should be highlighted. First, there are some concerns regarding the usefulness of salivary IgA as a biomarker in the detection of respiratory tract infection due to lack of reproducibility, low specificity, and sensitivity (61). Secondly, there are conflicting data in the literature regarding salivary IgA levels induced by exercise, with some studies reporting a decrease whilst others have reported an increase or no change (62). Thirdly, previous studies have reported a decrease in systemic IgA levels without leading to an increase in infection (63). Fourthly, exposure to pathogenic microbes may be reduced DR, possibly due to consumption of more fresh foods DR compared to other months (64). It is possible that oral health and microbial exposure from foods are poorer BR, which may explain the higher IgA levels detected in the saliva DR (65). In this context, a recent study involving mice reported that oral colonization by Candida albicans increases IgA production (65). Another study suggested that an increase in salivary IgA is an attempt by the immune system to counter the accumulation of microorganisms (64). Considering the aforementioned studies (63–65), the decrease in IgA levels DR may reflect a lower microbial colonization of the oral cavity DR. This is plausible since the number of hours when the mouth is exposed to foods and beverages is reduced DR compared to other periods when one considers the number of hours spent fasting and sleeping.

Overall, it seems that fluctuations in salivary parameters in Ramadan are not as significant as blood changes. These alterations are not enough to cause diseases in healthy participants. Nevertheless, we believe that further studies using other salivary biomarkers are needed in order to investigate correlations with the risk of oral disturbances or infections, such as caries, malodor, periodontal disease, or candidiasis in Ramadan.

In view of the absence of evidence about the impacts of RIF on oral health, we recommend the following four advices for people observing Ramadan: (i) adopt a well-balanced diet with sufficient hydration before Sahur and after Iftar; (ii) brush teeth, at least after Iftar and just after Sahur, before the dawn; (iii) rinse mouth without swallowing water for a better biofilm control and reduction of halitosis; and (iv) take care of the oral cavity, particularly for patients with chronic systemic diseases, especially with metabolic disorders (e.g., diabetes mellitus) in order to avoid the progression of a preexistent pathology (e.g., periodontal disease, dental caries). Finally, it is recommended that dentists carry out “dental procedures” with special precautions [e.g., administer intramuscular or trans-dermal treatment instead of oral agents] (36).

According to the JBI critical appraisal tool, precisely the checklist for cohort studies, the methodological quality is considered as “low.” In fact, no study succeeded to get the average score and items related to “confounding factors” and “sample size calculation.” Moreover, “salivary collection methods” were not reported in any of the six retained studies (Table 1). First, non-inclusion of a non-fasting control group can be considered as a “bias” since the variations in the assessed parameters cannot be exclusively attributed to RIF. However, it is important to note that including non-fasting participants is still problematic, due to religious considerations in Muslim countries. For that reason, the non-fasting control groups could be the participants themselves outside the Ramadan period (e.g., BR and/or AR). Given the circumstances of the Ramadan observance, and for practical reasons, the authors think that is more feasible and easier to control the parameters than to arrange a separate group of participants who do not observe Ramadan. Secondly, selecting participants by a convenience sample may be considered as a major confounding factor (66). Convenience sampling is a type of non-probability sampling methods based on the judgment of the investigator (66). Its low cost and comfort of use make it an easy choice for investigators. Nevertheless, it can lead to under/over representation of specific groups inside the sample (66). Thus, it may be impossible to make generalizations in the whole population. For these reasons, convenience sampling should be treated with caution. Thirdly, calculation of an optimal size is a crucial point since it helps avoid an inadequate power to detect statistical effects (67). Using few participants in a study may lead to lower “precision” in findings. A large sample size is, however, expensive and exposes more participants to procedures (67). Fourthly, the procedure of saliva collection was not well-described (Table 1). In fact, it is very important to standardize saliva sampling in order to make comparison between studies possible. Since saliva collection should be made at least one time DR, unstimulated saliva might be preferred. In fact, stimulated saliva must be collected by chewing sterile paraffin (68). Then, a minimum duration for sufficient saliva collection may be defined to ensure efficient analysis.

Additional limitations should be highlighted. For example, information about the season, the average ambient pressure, and/or the average ambient humidity was lacking in the included studies (Table 1). The average ambient temperature as well as the fasting duration were mentioned in some studies (26–29) (Table 1). Consequently, both climatic conditions and geographical locations strongly influence RIF (69). Also, the inclusion of patients with obesity (i.e., body mass index ≥ 30 kg/m²) may be considered as a limitation. In fact, a lower SFR was observed among obese compared to non-obese participants (70, 71). In addition, the inclusion of females and old participants could complicate the interpretation of saliva parameters (72, 73). Indeed, Mahesh et al. (72) reported significant changes in the pH and the buffer-capacity in post-menopausal females’ saliva compared to regularly menstruating ones. Besides, it is known that females do not fast all the month of Ramadan. Subsequently, the comparison with males may not be valid because they are not exposed to the same fasting period. With regard to age, changes in salivary pH, buffering-capacity, calcium, and proteins concentrations were reported (73). Finally, the number of evaluation sessions was heterogeneous. Therefore, saliva collection should be performed at least three times as follows: BR (e.g., 1 week BR), DR (e.g., during the last 7–10 days of Ramadan), and AR (e.g., 7–10 days AR). In future studies aiming to evaluate the effects of RIF on oral health, three important points should be reported. The first is related to the practice of fasting as a routine (e.g., some Muslims fast on Mondays and Thursdays during all the year). The inclusion of some participants who practice this ritual may influence some saliva parameters. The second point concerns the chewing stick, called “Miswak,” which is widely used in some Arab states of the Persian Gulf (74, 75). In fact, it seems that “Miswak” use increases SFR (76). The third point concerns the hydration status. The role of the hydration status BR and DR were not considered in the six retained studies and the differences observed in the concentrations of the different metabolites may be partly due to the hydration status, which can alter the salivary composition and SFR (77). The six studies involved in this review did not adjust the concentrations of the different salivary biomarkers before comparing the data obtained BR and DR. Therefore, in the future, it would be interesting to see if the differences observed are still present after adjusting the metabolites by factors, such as total protein content, saliva osmolality, SFR, and saliva secretion rate (78).

The critical limitation of this Systematic Review is our inability to make a strong clinical case for the impacts of RIF on saliva parameters. Table 4 summarizes some recommendations for designing future studies aiming to investigate the impacts of RIF on saliva parameters. It is recommended that researchers assess the antimicrobial, anticancer, and wound healing properties of fasting saliva (collected just before iftar) and compare it with non-fasting saliva. Moreover, it would be great to compare the fasting saliva proteome with the non-fasting saliva, and to see if the fasting saliva can be a source of novel peptides that display health benefits (79). This will address the “myth/superstition” in medieval Europe where fasting saliva was used as a medicine (80, 81).