Effect of probiotic colonization on oral hygiene and ventilator-associated pneumonia in ICU patients: a triple-blind randomized controlled trial

Abstract

Background

Mechanically ventilated patients are highly susceptible to poor oral hygiene and ventilator-associated pneumonia, mainly due to pathogenic bacterial colonization in the oral cavity. While antiseptic mouthwashes are commonly used, they may disrupt natural oral microbiota. Probiotics offer a potential alternative by promoting beneficial microbial colonization, enhancing oral hygiene, and reducing the risk of infection. This study aimed to evaluate the effect of probiotic colonization on oral hygiene and the incidence of ventilator-associated pneumonia in mechanically ventilated patients.

Methods

This triple-blind randomized controlled trial was conducted in the intensive care units (ICUs) of Shohada-ye Ashayer Hospital, Khorramabad, Iran. Eighty mechanically ventilated patients admitted to the ICU were randomly assigned to either an intervention or a control group using a simple randomization method. The intervention group received two daily doses of 250 mg of Streptococcus salivarius probiotic solution, administered orally for five consecutive days, while the control group received a placebo. Data collection tools included a demographic and clinical characteristics questionnaire, the Beck Oral Assessment Scale, and the Clinical Pulmonary Infection Score. Assessments were performed at baseline and on day six after the intervention. Data were analyzed using SPSS version 26, employing generalized linear models for comparative analysis of outcomes.

Results

Baseline demographic and clinical characteristics were comparable (p > 0.05), except for a lower Glasgow coma scale in the intervention arm (5.5 ± 1.3 vs. 6.4 ± 1.3, p = 0.003), which was adjusted for in subsequent analyses. Oral health status significantly improved in the intervention group compared with deterioration in controls (−1.82 ± 1.57 vs. +1.03 ± 1.66; adjusted F = 55.75, p < 0.001). Likewise, the clinical pulmonary infection score decreased in the intervention group but increased in the control group (−0.67 ± 1.29 vs. +1.35 ± 1.58; adjusted F = 30.27, p < 0.001). The proportion of patients at moderate to high risk of developing ventilator-associated pneumonia increased in the control group from 45% to 70%, whereas it decreased in the intervention group from 60% to 50%.

Conclusion

The findings of this randomized trial show that the use of Streptococcus salivarius was associated with improvement in oral hygiene scores and reduction in Clinical Pulmonary Infection Score values among mechanically ventilated ICU patients. These outcomes indicate that probiotic-based oral care may help enhance oral health and reduce the estimated risk of ventilator-associated pneumonia. Further research with broader samples can clarify the clinical impact of this approach.

Trial registration

The trial was registered in the Iranian Registry of Clinical Trials (IRCT20241019063429N1) in 2024-11-27.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-026-12711-1.

Introduction

Ventilator-associated pneumonia (VAP) is one of the most prevalent and severe hospital-acquired infections among critically ill patients admitted to intensive care units (ICUs) and undergoing mechanical ventilation [1]. Ventilator-associated pneumonia significantly increases morbidity, mortality, length of hospital stay, and healthcare costs [2]. The development of VAP is closely linked to bacterial colonization of the oral cavity and oropharynx, which serve as reservoirs for respiratory pathogens entering the lower respiratory tract [3].

Critically, patients are at particularly high risk due to factors such as impaired consciousness, frequent use of invasive devices, weakened immune defenses, and exposure to multidrug-resistant microorganisms [4]. According to the International Nosocomial Infection Control Consortium, infection-related mortality rates in ICUs remain considerably high, especially in Middle Eastern countries [5]. In Iran, studies report that approximately 30% of ICU patients experience hospital-acquired infections, with VAP representing one of the leading causes [6, 7].

Maintaining oral hygiene is critical for preventing respiratory infections, especially in mechanically ventilated patients. Due to reduced salivary flow, medication-induced xerostomia, and impaired swallowing, ICU patients experience rapid changes in their oral microbiota, which can lead to biofilm formation and colonization by pathogenic species [8]. Dental plaque acts as a persistent reservoir for respiratory pathogens, increasing the risk of microaspiration and VAP [9]. Scientific evidence supports a strong association between poor oral hygiene and VAP incidence [10]. For example, Karimi et al. reported that adherence to standardized oral care protocols significantly reduced VAP rates in mechanically ventilated patients [11].

Despite its importance, oral hygiene is often neglected in ICUs due to limited nursing resources, lack of training, and prioritization of life-threatening conditions over preventive care [12]. Studies have shown that more than 80% of ICU patients require dental interventions, including oral hygiene procedures or invasive dental treatments, yet these needs are frequently unmet [13]. Without proper oral care, colonization of the oral cavity by respiratory pathogens accelerates, increasing the risk of nosocomial infections and poor clinical outcomes [14, 15].

Current preventive strategies for VAP include semi-recumbent positioning, head-of-bed elevation, suctioning of secretions, and strict adherence to infection control measures. However, effective oral hygiene remains a cornerstone intervention [16]. Mechanical plaque removal, antiseptic rinses, and standardized oral care protocols significantly reduce the bacterial burden and, consequently, the risk of VAP [17]. Among chemical interventions, chlorhexidine mouthwash is widely used due to its proven efficacy against oral pathogens and ability to reduce hospital-acquired respiratory infections [18, 19]. Nonetheless, prolonged chlorhexidine use can cause mucosal irritation, tooth staining, taste changes, and disruption of the oral microbiome, limiting its long-term use in critically ill patients [20, 21]. These limitations highlight the need for safer and more effective alternatives.

Probiotics have emerged as a promising adjunctive strategy in maintaining oral health and preventing VAP [22]. Probiotics are live microorganisms that provide health benefits by competing with pathogens, producing antimicrobial substances, and modulating the immune system [23, 24]. Among various strains, Streptococcus salivarius (S.salivarius) has been particularly investigated for its role in oral health due to its ability to inhibit colonization by respiratory pathogens [25, 26].

Given the high prevalence of VAP, its associated mortality, and the limitations of existing preventive measures, evaluating novel strategies is a clinical priority. Probiotic oral care is a safe, non-invasive, and potentially cost-effective approach that may enhance oral hygiene, restore microbial balance, and reduce the risk of respiratory complications in mechanically ventilated ICU patients [8].

Therefore, the present study aimed to investigate the effect of probiotic-based oral care with S. salivarius on oral hygiene status and the incidence of VAP in critically ill patients undergoing mechanical ventilation. The findings of this research may contribute to establishing evidence-based protocols for infection prevention, improving nursing practices, and enhancing patient outcomes in intensive care settings.

Methods

Study design

This study was a randomized, placebo-controlled, triple-blind clinical trial conducted in accordance with the CONSORT guidelines from December 2024 to August 2025.

Participants

The study population included patients admitted to the ICUs of Shohada Ashayer Educational and Medical Center in Khorramabad, Iran. Eligible participants were those who met the inclusion and exclusion criteria. Inclusion criteria were: Age 18–85 years; Level of consciousness with Glasgow Coma Scale (GCS) < 10; Presence of an endotracheal tube, and within 48 hours of intubation. Exclusion criteria were: Development of unexpected complications such as allergic reactions or gastrointestinal problems; Withdrawal of consent by the patient’s legal guardian; Removal of the ETT during the trial; Use of oral rinses or antifungal agents during the study; Patient death during the study; Pre-existing pneumonia diagnosis; Presence of oral trauma or wounds; Use of dentures; Signs of pre-existing oral diseases.

Sample size

The sample size was determined using the Beck Oral Assessment Scale (BOAS) as the primary outcome measure. The following statistical assumptions guided the calculation: a significance level of α = 0.05, a statistical power of 90% (β = 0.10), and an expected mean difference (δ) of 2 points between the intervention and control groups. The pooled standard deviation was estimated at σ = 2.5, based on prior clinical data and preliminary observations [27]. Using these assumptions, the standardized effect size was calculated as δ/σ = 0.80, representing a clinically meaningful improvement in oral status. The sample size was estimated using the formula for comparing two independent means, yielding a requirement of 33 participants per group. To account for a potential 20% dropout rate, the total sample size was increased to 40 participants per group, resulting in an overall sample of 80 patients.

Randomization

Eligible patients were randomly assigned to either the intervention or the control group using simple randomization. Each participant was assigned to a group based on a randomly generated sequence of letters A and B, with A representing the intervention group and B the control group. The sequence was generated using a random number table, with a single digit between 0 and 9 selected for each assignment. If the digit was between 0 and 4, the patient was assigned to group A; if the digit was between 5 and 9, the patient was assigned to group B. This process continued until all participants were allocated to one of the two groups.

Blinding

This was a triple-blind clinical trial in which patients, sample collectors, caregivers, and data analysts were all blinded to group allocation. The probiotic and placebo sachets were prepared and coded (A and B) by the manufacturing company, with identical appearance, taste, and packaging to ensure allocation concealment. The allocation codes remained inaccessible to the research team throughout data collection and analysis. Code-breaking was permitted only in the event of a serious adverse event requiring unblinding for clinical decision-making as an emergency measure; however, no such event occurred, and codes were disclosed solely after completion of statistical analysis.

Interventions

The intervention group received S. salivarius (250 mg, 10⁹ CFU) twice daily at 06:00 and 18:00 for five consecutive days. The dosage regimen was based on the protocol of Golparvar et al. (2019) [28]. The probiotic used in this study was based on the commercial oral preparation Lactogam®, formulated by Zist Takhmir Pharmaceutical Company (Tehran, Iran). Lactogam® contains two specific oral probiotic strains, S.salivarius K12 and M18, formulated to support oral and pharyngeal health. For this study, we intentionally selected a single probiotic product containing these two S. salivarius strains, rather than a multi-strain mixture. This approach allowed for a more precise evaluation of the strain-specific effects, minimized potential inter-strain interactions, and facilitated clearer interpretation of both clinical outcomes and safety.

To prepare the intervention, lozenges were crushed into powder and dissolved in 5 mL of distilled water immediately before administration. The resulting suspension was applied to the oral cavity using a sterile swab after the mouth was rinsed with saline and gently brushed. No rinsing was performed afterward to support colonization.

The placebo consisted of lozenges identical in appearance, weight, packaging, and administration method, but lacking viable probiotic microorganisms. Placebo tablets were similarly crushed and dissolved in 5 mL of distilled water before oral application.

Instruments

Demographic and baseline characteristics were recorded using a structured questionnaire. These included age, sex, marital status, education level, admission diagnosis, route of nutrition, level of consciousness (GCS), duration of intubation, and gastrointestinal condition. Disease severity was assessed using the Acute Physiology and Chronic Health Evaluation II (APACHE II) score. In addition, data on antibiotic therapy, including the number of antibiotics prescribed and their spectrum, were collected (Supplementary file 1).

Oral health status was measured using the BOAS, an observational checklist consisting of five domains (lips, mucosa and gingiva, teeth, tongue, and saliva). Each item is scored from 1 to 4, yielding a total score between 5 and 20, with higher scores indicating poorer oral health. The validity and reliability of this instrument were previously confirmed by Tu et al. [29], and in Iran by Sefrabadi et al., who reported a test–retest reliability of 0.92 [30].

The incidence of VAP was evaluated using the Clinical Pulmonary Infection Score (CPIS), initially developed by Pugin et al. [31] and subsequently validated in Iran by Yaghobi et al. (2019) [32]. CPIS is based on six readily available clinical and laboratory parameters: (1) body temperature, (2) leukocyte count, (3) tracheal secretions, (4) chest radiography, (5) microbiology of endotracheal aspirates, and (6) oxygenation (PaO₂/FiO₂ ratio). Each variable is scored from 0 to 2, producing a total score of 0–12. A score of 6 or higher indicates a high probability of VAP. To ensure consistent evaluation, each of the six CPIS components was assessed as follows: Body temperature: measured three times daily (morning, afternoon, evening) using a mercury axillary thermometer. According to CPIS criteria, values <36 °C or >38.5 °C scored 1, while <35 °C or >39 °C scored 2; values between 36 and 38.4 °C scored 0. Leukocyte count: daily venous blood samples were analyzed using CBC with differential. Counts > 11,000 or <4,000 cells/mm³ scored 1; if accompanied by a left shift, a score of 2 was assigned. Normal values scored 0. Tracheal secretions: assessed before and after the intervention by two ICU nurses for volume, color, viscosity, and odor. Purulent or foul-smelling secretions scored 2; colored or moderately abundant secretions scored 1; clear, scant secretions scored 0. Chest radiography: Portable chest X-rays were obtained and interpreted by ICU physicians or radiologists without knowledge of the patient’s condition. New bilateral or diffuse infiltrates were scored 2, new localized infiltrates were scored 1, and no new findings or chronic changes were scored 0. Microbiological cultures: endotracheal aspirates were collected aseptically and cultured for common respiratory pathogens (Pseudomonas, Acinetobacter, Klebsiella, etc.). Positive cultures scored 1; negative or non-significant growth scored 0. Oxygenation: PaO₂/FiO₂ ratio was calculated from arterial blood gas analysis and ventilator settings. Ratios ≤ 240 scored 2, 241–300 scored 1, and >300 scored 0 [32].

Statistical methods

Descriptive statistics included frequency distributions, bar and pie charts, means ± standard deviation, or medians with interquartile ranges. Normality was assessed using the Shapiro–Wilk test. Independent t-tests or Mann–Whitney U tests were used for two-group comparisons, while one-way ANOVA or Kruskal–Wallis tests were used for multi-group analyses, along with appropriate post-hoc tests. Multivariate analyses were performed using analysis of covariance (ANCOVA) and marginal longitudinal models with a logit link function. A two-sided p-value < 0.05 was considered statistically significant. Analyses were performed using IBM SPSS Statistics version 26.

Results

Eighty patients were randomly assigned to one of the two study arms (n = 40 per arm). All randomized participants completed the study and were included in the analyses, adhering to the intention-to-treat principle (n = 80). There were no missing outcome data for the primary or secondary endpoints (Fig. 1). Normality of continuous variables was evaluated using the Kolmogorov–Smirnov and Shapiro–Wilk tests (p > 0.05). The baseline demographic and clinical characteristics of the two groups are presented in Table 1. Groups were similar across most baseline variables (p > 0.05). The mean GCS at enrollment differed between groups: control, 6.4 ± 1.3, versus intervention, 5.5 ± 1.3 (p = 0.003) (Table 1). Because GCS differed significantly at baseline, it was entered as a covariate in later adjusted analyses.

Fig. 1.

CONSORT flow diagram

Table 1.

Baseline demographic and clinical characteristics (n = 80)

Characteristic	Control	Intervention	p-value
Age, mean (SD), years	46.3 (14.7)	46.1 (18.5)	0.912
GCS, mean (SD)	6.4 (1.3)	5.5 (1.3)	0.003
Duration of intubation, mean (SD), days	2.6 (0.6)	2.5 (0.7)	0.255
Sex, n (%)
Female	17 (42.5%)	19 (47.5%)	0.653
Male	23 (57.5%)	21 (52.5%)
Marital status, n (%)
Without spouse (divorced/widowed)	10 (25.0%)	7 (17.5%)	0.714
Married	19 (47.5%)	21 (52.5%)
Never married	11 (27.5%)	12 (30.0%)
Education, n (%)
Illiterate	3 (7.5%)	9 (22.5%)	0.091
Primary/middle	9 (22.5%)	4 (10.0%)
Secondary/diploma	15 (37.5%)	10 (25.0%)
University	13 (32.5%)	17 (42.5%)
Admission diagnosis, n (%)
Trauma	18 (45.0%)	18 (45.0%)	0.308
Surgical	11 (27.5%)	5 (12.5%)
Medical (internal)	7 (17.5%)	10 (25.0%)
Poisoning	4 (10.0%)	7 (17.5%)
Nutrition route, n (%)
NPO	21 (52.5%)	19 (47.5%)	0.655
Gavage	19 (47.5%)	21 (52.5%)
Severity of illness, n (%)
APACHE II ≤ 30	21 (52.5%)	19 (47.5%)	0.655
APACHE II > 30	19 (47.5%)	21 (52.5%)
Number of antibiotics, mean (SD)	2.1 (0.8)	2.3 (0.9)	0.412
Spectrum of antibiotics, n (%)
Narrow spectrum	14 (35.0%)	12 (30.0%)	0.527
Broad spectrum	26 (65.0%)	28 (70.0%)

Note: Between-group tests: independent-samples t-test or Mann–Whitney U for continuous variables; χ² test for categorical variables. SD: Standard Deviation; GCS: Glasgow Coma Scale; APACHE II: Acute Physiology and Chronic Health Evaluation II

Table 2 summarizes group means and changes for the BOAS and the CPIS at baseline and at study end (day 6). To account for baseline imbalance in GCS and BOAS, the effect of group assignment (intervention vs. control) on BOAS change was analyzed using ANCOVA, with baseline BOAS, education, and baseline GCS entered as covariates. After adjustment, group assignment showed a large, statistically significant effect on BOAS change, indicating that patients in the probiotic group experienced a greater improvement in oral health than controls (adjusted group effect: F = 55.75, p < 0.001). Baseline BOAS was also a strong predictor of the final score (F 14.84, p < 0.001), whereas baseline GCS (p = 0.832) was not a significant predictor (Table 3).

Table 2.

Mean scores (baseline, end of study) and within-group change for BOAS and CPIS

Outcome	Time point	Control	Intervention
BOAS, mean (SD)	Baseline	8.50 (2.33)	9.17 (2.15)
	End of study (day 6)	9.35 (2.62)	7.35 (1.39)
	Change (end – baseline)	+1.03 (1.66)	−1.82 (1.57)
CPIS, mean (SD)	Baseline	5.55 (2.59)	6.13 (2.81)
	End of study (day 6)	6.90 (2.60)	5.45 (2.43)
	Change (end – baseline)	+1.35 (1.58)	−0.67 (1.29)

Note: SD: Standard Deviation; BOAS: Beck Oral Assessment Scale; CPIS: Clinical Pulmonary Infection Score

Table 3.

Results of ANCOVA for the effects of intervention group, baseline scores, and baseline GCS on changes in BOAS and CPIS

Source	df	MS	F	p-value
BOAS
Model (corrected)	7	30.77	14.79	<0.001
Intercept	1	0.37	0.18	0.677
Group (intervention)	1	115.98	55.75	<0.001
Baseline BOAS	1	30.87	14.84	<0.001
Baseline GCS	1	0.09	0.05	0.832
Error	72	2.08
CPIS
Model (corrected)	6	18.48	10.14	<0.001
Intercept	1	9.16	5.02	0.028
Group (intervention)	1	55.16	30.27	<0.001
Baseline CPIS	1	24.40	13.39	<0.001
Baseline GCS	1	0.25	0.14	0.710
Error	73	1.82

Note: The table presents the key test statistics used to evaluate the adjusted effect of intervention on each outcome. The ANCOVA analyses were adjusted for baseline outcome scores (baseline BOAS or baseline CPIS) and baseline GCS. The degrees of freedom reported reflect the model and residual degrees of freedom as described above. df: Degrees of Freedom; MS: Mean Square; F: F-statistic; BOAS: Beck Oral Assessment Scale; GCS: Glasgow Coma Scale; CPIS: Clinical Pulmonary Infection Score

Change in CPIS (end minus baseline) was analyzed similarly using ANCOVA, controlling for baseline CPIS, education, and baseline GCS. The results demonstrated a highly significant group effect on CPIS change, confirming the intervention’s favorable impact on reducing the risk of pulmonary infection (adjusted group effect: F = 30.27, p < 0.001). Baseline CPIS was also a significant predictor (F = 13.39, p < 0.001). The baseline GCS was not a significant predictor after adjustment (p = 0.710) (Table 3).

Descriptively, the distribution of VAP risk categories moved in opposite directions in the two arms. The proportion of patients categorized as having a moderate/high (score≥6) probability of VAP increased in the control arm from 45% at baseline to 70% at day 6. In contrast, in the intervention arm, the proportion with a moderate to high probability decreased from 60% at baseline to 50% at day 6 (Table 4).

Table 4.

Distribution of patients by estimated probability of VAP (low vs. moderate/high) at baseline and day six

Timepoint	Control—low (n, %)	Control—moderate/high (n, %)	Intervention—low (n, %)	Intervention—moderate/high (n, %)
Baseline	22 (55%)	18 (45%)	16 (40%)	24 (60%)
Day 6	12 (30%)	28 (70%)	20 (50%)	20 (50%)

Note: The cutoff point of ≥6 on the CPIS was used to define the presence of ventilator-associated pneumonia (VAP). n: Number; %: Percentage

No serious or non-serious adverse events related to the probiotic intervention were observed during the study period. All participants tolerated the probiotic and routine ICU care without any reported complications, and no withdrawals occurred due to safety concerns.

Discussion

The present study demonstrated that probiotic-based oral care contributed to meaningful improvements in oral health status and lowered clinical pulmonary infection score among critically ill patients. The two groups were generally comparable in terms of demographic and baseline clinical characteristics, which increases confidence that the observed differences are attributable to the intervention rather than to pre-existing imbalances. Although the intervention group had lower baseline consciousness, statistical adjustment for this factor addressed this factor, and the overall comparability of the groups supports the validity of the findings.

The observed improvement in oral health status, despite a worse baseline BOAS in the intervention group, supports a genuine intervention effect rather than a simple regression-to-the-mean phenomenon. Baseline BOAS was a significant predictor of outcome, and the intervention effect remained strong after adjustment for other variables. Mechanistically, critically ill patients are prone to rapid dysbiosis of the oral microbiota, biofilm maturation, hyposalivation, and impaired mechanical clearance—processes that accelerate deterioration of oral status within 48–72 hours [33]. Probiotic strains, such as S. salivarius, can compete for adhesion sites, produce inhibitory substances, and modulate local host responses, thereby reducing pathogen load and stabilizing the oral ecosystem. Clinical studies and meta-analytic evidence support the beneficial effects of probiotics on periodontal and oral-microbial indices [33–35]. Our findings therefore align with trials showing improved oral indices after probiotic interventions and highlight that a protocolized, multi-component oral care regimen, including mechanical cleaning, probiotic instillation, and moisture management, can reverse an otherwise progressive decline in oral health in ventilated patients.

The reduction in CPIS and the favorable shift in estimated VAP probability in the intervention arm are biologically plausible and clinically significant. VAP most often results from microaspiration of oropharyngeal secretions and from biofilm formation on the endotracheal tube; the loss of natural defense mechanisms in intubated patients facilitates transfer of oral pathogens to the lower respiratory tract [33, 34]. Meta-analyses and systematic reviews report that probiotic interventions can be associated with reductions in VAP incidence in selected populations [36, 37]. However, not all trials are concordant, and effect estimates vary depending on the population, probiotic strain, and study quality [28, 38]. Importantly, recent syntheses have highlighted heterogeneity in trial designs and overall low–to–moderate certainty of evidence, so conclusions must be tempered [39].

Several randomized controlled trials and meta-analyses have investigated the use of probiotics for preventing VAP, but the findings remain mixed. While studies by Simpos et al. [40], Van Roijen et al. [41], and Su et al. [42] reported significant reductions in VAP incidence, others, such as Gu et al. [43] and Wang et al. [44], found no statistically significant benefits. These inconsistencies may relate to variations in patient populations, probiotic strains, dosages, and delivery methods, underscoring the need for further well-designed clinical studies.

In our study, objective reductions in CPIS and a possible decrease in VAP risk in the probiotic arm, together with a significant adjusted group effect, support the hypothesis that targeted oral microbiome modulation can reduce surrogate and clinical indicators of ventilator-associated pulmonary infection [31, 32].

If confirmed in larger and more diverse settings, the present results suggest that adding probiotic-based oral care to ICU oral care bundles could be an effective, low-risk adjunct to standard infection prevention practices. Beyond the direct oral health benefit, reduced CPIS has potential downstream effects on antibiotic exposure, ventilator days, and ICU length of stay; several studies and pooled analyses have reported reductions in antibiotic days and ICU length of stay associated with probiotic or synbiotic strategies [42, 45, 46]. Compared with chemical antiseptics, such as chlorhexidine, which can rapidly reduce bacterial load but is associated with mucosal irritation, taste disturbance, and microbiome perturbation [18, 19, 21], probiotic approaches may offer a gentler, ecology-preserving adjunct or alternative. Implementation, however, requires attention to strain selection, reliable manufacturing, dosing schedules, staff training, and safety monitoring in severely immunocompromised patients.

To build a stronger evidence base, future studies should be multicenter, adequately powered for clinically meaningful endpoints (ventilator-free days, antibiotic consumption, ICU and hospital length of stay, and mortality), and should compare probiotic oral care head-to-head with established antiseptic regimens (e.g., chlorhexidine) or combined strategies. Trials must standardize probiotic strain, dose, and delivery method and should incorporate mechanistic substudies to document ecological effects and colonization durability. Given the heterogeneity and variable quality of existing evidence, systematic harmonization of baseline measures and reporting (as urged by prior reviews) is essential [39, 47]. Pragmatic cost-effectiveness analyses and safety surveillance, particularly in the context of concurrent broad-spectrum antibiotic exposure, are also recommended.

Key strengths of this trial include a randomized, triple-blinded design, use of validated instruments for oral health (BOAS) and pulmonary infection assessment (CPIS), and complete outcome ascertainment with no losses to follow-up. The ANCOVA approach accounted for the baseline imbalance in GCS and baseline outcome values, strengthening causal inference.

Despite the strengths of this study, several limitations should be acknowledged. The investigation was conducted at a single center with a relatively brief intervention period of five days, which may constrain the external validity of the findings and preclude evaluation of long-term outcomes. Although baseline differences in consciousness were analytically adjusted for, residual confounding may remain. Comprehensive sequencing of the oral and respiratory microbiomes was not performed, leaving the precise ecological effects of S.salivarius in this patient population undefined. Additionally, the generalizability of the results is influenced by factors such as strain specificity, dosage, and local manufacturing quality, limiting the applicability of the findings to the specific preparation and regimen used in this study. Another important limitation is the heterogeneity of antibiotic therapy among ICU patients. Variations in both the number of antibiotics administered and their spectrum of activity, ranging from broad- to narrow-spectrum agents, may have affected outcomes. Although statistical adjustments were made to account for these factors, residual confounding cannot be ruled out entirely.

Conclusion

This study demonstrated that probiotic administration improved oral hygiene and reduced CPIS scores in critically ill patients undergoing mechanical ventilation, suggesting a beneficial effect on oral health and the estimated risk of VAP. Continued research with larger and multicenter populations, longer follow-up, and microbiological confirmation is recommended to determine the full clinical value and long-term outcomes of probiotic-based oral care.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

VAP

Ventilator-associated pneumonia

ICU

Intensive Care Unit

GCS

Glasgow Coma Scale

APACHE II

Acute Physiology and Chronic Health Evaluation II

BOAS

Beck Oral Assessment Scale

CPIS

Clinical Pulmonary Infection Score

Author contributions

A.S.: Study design, sampling, manuscript writing; F.E.: Study design, statistical analysis, manuscript writing; A.B.: Study design, analysis of chest radiographs and tests, manuscript writing; M.H.: Sampling and manuscript writing; S.Y.: Idea generation, study design, manuscript editing, supervisor.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The datasets generated and analyzed during the current study are not publicly available due to patient confidentiality; however, they are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The Vice-Chancellor approved the study protocol for Research and Technology of Lorestan University of Medical Sciences and the Research Ethics Committee of Lorestan University of Medical Sciences, Khorramabad, Iran (Approval code: IR.LUMS.REC0.1403.329). Written and oral informed consent was obtained from the legal guardians of all participants prior to enrollment. All procedures were conducted in accordance with the ethical standards of the institutional research committee, the principles of the Declaration of Helsinki (1964) and its later amendments, with strict attention to patient confidentiality. Ethical principles, including respect for dignity, autonomy, justice, and beneficence, were fully observed throughout the study. The clinical care and routine medications of patients were neither withheld nor altered, and participants’ safety and well-being were prioritized at all times. The trial was discontinued immediately if consent was withdrawn or the patient’s condition deteriorated.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Use of artificial intelligence

Artificial intelligence tools (ChatGPT, OpenAI) were used solely to assist with English translation, language polishing, and figure preparation (one CONSORT diagram). They were not used for data analysis, statistical interpretation, or the generation of scientific content. All intellectual responsibility for the content remains with the authors.