Impact of the BioFire^® FilmArray^® Gastrointestinal Panel on clinical management of adult acute gastroenteritis: a retrospective analysis from a Saudi tertiary-care setting

Abstract

Background

Acute gastroenteritis (AGE), a leading cause of worldwide morbidity, can be triggered by a wide array of pathogens. Syndromic molecular platforms, such as the BioFire^® FilmArray^® Gastrointestinal Panel (GIP), allow rapid and comprehensive detection of enteric pathogens; however, evidence on their utility from Saudi Arabia is lacking. This study therefore aimed to assess the prevalence of enteric pathogens, determine the detection yield of the FilmArray GIP, and explore its potential impact on clinical management.

Methods

We conducted a retrospective cohort study at Dr. Soliman Fakeeh Hospital, Jeddah, Saudi Arabia, from January 2023 to December 2024. All adult patients (≥ 18 years) with suspected AGE tested using the FilmArray GIP were eligible for inclusion. Clinical and laboratory data, GIP findings, antimicrobial therapy, and clinical outcomes were extracted from electronic medical records.

Results

A total of 176 adult patients with AGE were included in this analysis. The FilmArray GIP detected bacterial pathogens in 86.4% of patients, while viral and parasitic agents were identified in 20.5% and 10.2% of the cohort, respectively. Co-detections were observed in 46.6% of patients, although their clinical relevance remains uncertain. Stool microscopy failed to detect any parasitic pathogens, whereas the multiplex panel identified parasites in 18 patients (p < 0.001). Detection of Salmonella was more than five times higher with the FilmArray GIP compared with stool culture (18.2% versus 3.4%; p < 0.001). The FilmArray GIP was associated with antimicrobial modifications, with no antibiotics prescribed in 30.1% of patients, whereas targeted antimicrobials were initiated in 27.3%. Notably, 8.5% of cases required transmission-based isolation, with Clostridioides difficile infections were more frequently associated with isolation compared with norovirus cases (42.3% versus 15.4%; p = 0.11).

Conclusions

The FilmArray GIP demonstrated higher detection yield for enteric pathogens compared with conventional diagnostic techniques. Rapid identification of enteric pathogens was temporally associated with antimicrobial optimization and infection control decisions. Taken together, these findings highlight the potential role of syndromic molecular testing in the management of adult AGE.

Clinical trial number

Not applicable.

Introduction

Acute gastroenteritis (AGE) is a common infectious disease characterized by inflammation of the stomach and intestines, leading to symptoms such as nausea, vomiting, diarrhea, and abdominal cramps [1]. In contrast to persistent GE that extends beyond 14 days but resolves within 30 days, acute infection usually lasts for a period of less than 14 days [2]. While most cases are self-limiting, AGE remains a public health concern, due to its association with substantial morbidity, mortality, and healthcare costs [3]. According to the latest global burden of diseases (GBD) estimates, diarrhea emerged as the world’s second most prevalent illness, with 467 million new cases reported, trailing only upper respiratory tract infections [4].

The etiology of AGE is miscellaneous, including bacterial, viral, and parasitic culprits. Current evidence indicates that bacterial pathogens, such as enteropathogenic Escherichia coli (EPEC), enteroaggregative E. coli (EAEC) and Clostridioides difficile (Cl. difficile) are primary drivers of adult AGE [5]. Several viruses, including noroviruses and rotaviruses are also commonly implicated, with noroviruses infecting all age groups, while rotaviruses mainly affect infants and young children. Other viruses, including adenoviruses, astroviruses, and sapoviruses are well-known causes of enteric infections [6]. Parasitic agents, although less frequent than bacterial or viral pathogens, have been reported in adults with diarrheal illness, particularly Giardia lamblia and Cryptosporidium species [7]. According to a recent retrospective study from Al-Ahsa, Saudi Arabia, bacterial, viral, and parasitic agents were identified in 50.5%, 24.8%, and 3.8% of the examined samples, respectively [8].

Precise and prompt identification of the causative agents of AGE is fundamental for judicious use of antimicrobials, applying appropriate infection control strategies, and achieving optimal patient outcomes, particularly in the era of rising antimicrobial resistance [9]. Traditionally, the laboratory diagnosis of AGE relies on stool microscopy, culture-based methods, and immunoassays. However, these techniques have limited specificities and sensitivities and lack the ability of synchronous detection of multiple pathogens [10]. Over the past few years, multiplex polymerase chain reaction (PCR) panels have revolutionized the diagnosis of infectious diseases. One such platform is the BioFire^® FilmArray^® Gastrointestinal Panel (GIP; BioFire Diagnostics Inc., USA), which provides syndromic identification of 22 common gastrointestinal pathogens directly from stool samples within approximately one hour [11].

Despite growing evidence of the diagnostic potential of the FilmArray GIP, data from Saudi Arabia on the use of multiplex molecular testing in AGE remain limited. To address this gap, we conducted this retrospective study to evaluate (i) the prevalence of enteric pathogens associated with adult AGE using the FilmArray GIP, (ii) the agreement between the FilmArray GIP and conventional diagnostic techniques, and (iii) modifications in clinical management and infection control measures following the FilmArray GIP results.

Patients and methods

Ethics approval and consent statement

The study followed the ethical standards of the institutional and national research committee and the tenets of the 1964 Declaration of Helsinki and its later amendments. The study protocol was reviewed and approved by the Institutional Review Board (IRB) of Fakeeh Care Group (reference number FIRB-25-0033). All data were anonymized before analysis, and patient confidentiality was strictly maintained throughout the study. The requirement for informed consent was waived due to the retrospective nature of the study and the use of de-identified dataset.

Study design and setting

Between January 2023 and December 2024, this single-center, retrospective cohort study was conducted at Dr. Soliman Fakeeh Hospital (DSFH), a private, 500-bed tertiary healthcare center in Jeddah, Saudi Arabia. The hospital serves as a referral center and delivers both inpatient and outpatient tertiary care to residents across the Kingdom.

Study population

Inclusion criteria

All adult patients (≥ 18 years) who underwent stool testing using the BioFire^® FilmArray^® GIP for suspected AGE during the study period were eligible for inclusion. AGE was defined as passage of three or more loose or watery stools per day for less than 14 days, with or without nausea, vomiting, abdominal pain, or fever [12].

Exclusion criteria

Patients were excluded if they met any of the following criteria: (1) age < 18 years, (2) non-infectious diarrhea (e.g., inflammatory bowel disease), (3) persistent diarrhea (≥ 14 days), and (4) incomplete clinical or laboratory data. For patients who submitted several stool samples during the study period, only the first sample collected was included in our analysis to ensure one sample per patient.

Data extraction of the study cohort

We conducted a retrospective review of DSFH electronic medical and laboratory records of all adult patients diagnosed with AGE using the FilmArray GIP during the study period. Eligible patients were primarily identified using the FilmArray GIP laboratory service number, after which patient medical record numbers (MRNs) were used to extract the relevant data to the study cohort. Patient demographics, body mass index (BMI; kg/m²), comorbidities, and predisposing factors for AGE (e.g., consumption of contaminated food, history of recent travel, and recent use of broad-spectrum antibiotics) were abstracted. Further data collected included the date of AGE episode, presenting symptoms (nausea, vomiting, diarrhea, abdominal pain, and fever), duration of illness prior to presentation, complications (dehydration, hospitalization, intensive care unit [ICU] admission, septic shock), and mortality.

Laboratory data were retrieved from the hospital electronic medical records, including total white blood cell (WBC) count (×10³/µl), differential WBC count (%), and C-reactive protein (CRP; mg/L). Results of the FilmArray GIP were obtained directly from the microbiology laboratory information system. Stool microscopy and culture were routinely done for all included patients, and their results were also collected.

Data related to patient management after availability of the FilmArray GIP results were extracted from the hospital electronic medical records to assess the clinical utility of the panel. These included modifications in antimicrobial therapy and implementation of transmission-based isolation precautions. Importantly, decisions related to antimicrobial prescribing were explicitly documented in daily progress notes as being based on the FilmArray GIP results, in conjunction with clinical evaluation by the treating clinicians.

Definitions

Empiric therapy was defined as the initiation of broad-spectrum antimicrobial treatment before microbiological confirmation based on the most probable pathogens. Targeted therapy referred to adjustment of antimicrobial treatment according to available culture and susceptibility results, and was supported by findings from the FilmArray GIP for early clinical decision-making. Escalation described the broadening of antimicrobial coverage in response to clinical deterioration or insufficient initial therapy, while de-escalation involved narrowing or discontinuation of antimicrobial agents once clinical and microbiological data warranted [13].

Statistical analysis

Data were analyzed using the Statistical Package for Social Sciences (SPSS) software (IBM Corp., IBM SPSS Statistics for Windows, Version 28.0, Armonk, NY, USA). Continuous variables were assessed for normality using the Shapiro–Wilk test. Normally distributed continuous variables were expressed as mean ± standard deviation (SD), while non-normally distributed variables were presented as median and interquartile range (IQR). The independent samples t-test or Mann–Whitney U test was used to compare continuous variables between two independent groups, as appropriate. The Pearson’s chi-squared (χ²) test was applied to assess associations between categorical variables, while Fisher’s exact test was used when more than 20% of the expected cell counts were less than five. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated to measure the strength of any associations. Agreement between the FilmArray GIP and conventional diagnostic methods was assessed using overall percent agreement, defined as the proportion of concordant results across all tested samples. Statistical significance was set at p-values < 0.05 (2-tailed).

Results

Demographic and clinical characteristics of included patients

During the study period, 307 patients underwent stool testing using the FilmArray GIP. Of these, 37 were excluded for not meeting the inclusion criteria (17 were children, 9 had non-infectious diarrhea, 7 had persistent diarrhea, and 4 had incomplete data). In addition, 94 patients had negative FilmArray GIP result. A total of 176 patients were included in the final analysis. The mean age of the study population was 44.9 ± 19.7 years (range, 18–94 years), with a slight male predominance (54% versus 46%). Most patients presented to the emergency department (92%), while outpatient visits accounted for 8%. The mean duration of illness was 3.6 ± 3.2 days (range, 1–13 days). Diarrhea was reported in all patients, including watery diarrhea in 20.5% and bloody diarrhea in 6.8%. Abdominal pain was the most frequently associated symptom (88.6%), followed by vomiting (71.6%), whereas fever was the least common presentation (Table 1).

Table 1.

Baseline demographic and clinical characteristics of the study cohort

Variables	Participants (n = 176)
Age (years)
Mean ± SD (Min–Max)	44.9 ± 19.7 (18–94)
Gender, n (%)
Male	95 (54.0%)
Female	81 (46.0%)
BMI (kg/m2)
Mean ± SD (Min–Max)	29.6 ± 7.1 (16.3–53.6)
Visit type, n (%)
ED	162 (92.0%)
OPD	14 (8.0%)
Duration of illness (days)
Mean ± SD (Min–Max)	3.6 ± 3.2 (1–13)
Major presenting symptom, n (%)
Diarrhea	176 (100.0%)
Watery diarrhea	36 (20.5%)
Bloody diarrhea	12 (6.8%)
Abdominal pain	156 (88.6%)
Vomiting	126 (71.6%)
Fever	72 (40.9%)
Character of abdominal pain, n (%)
Diffuse	126 (80.8%)
Lower abdominal	6 (3.8%)
Periumbilical	2 (1.3%)
Colicky	20 (12.8%)
Pain score
Mean ± SD (Min–Max)	3.3 ± 2.1 (0–10)

Data are presented as numbers (%) for categorical variables and as mean ± SD for continuous variables. SD, standard deviation; Min, minimum; Max, maximum; BMI, body mass index; ED, emergency department; OPD, outpatient department

Comorbidities, potential risk factors, and laboratory findings of included patients

Chronic comorbidities were commonly observed among the study cohort, with diabetes mellitus (29.5%) and hypertension (26.1%) being the most prevalent, followed by chronic kidney disease (11.4%). Reported potential risk factors were exposure to contaminated food or water (8%) and recent travel history (6.8%). Exposure to broad-spectrum antibiotics was recorded in 2.3% of cases. Laboratory results showed a median CRP level of 22.1 mg/L (IQR: 7.2–74.5). The median total WBC count was 7.9 × 10³/µl (IQR: 5.3–12.8), with neutrophil predominance (73.9%) and a low relative lymphocyte count (Table 2).

Table 2.

Comorbidities, risk factors, and laboratory findings of the study population

Variables	Participants (n = 176)
Comorbidities, n (%)
Diabetes mellitus	52 (29.5%)
Hypertension	46 (26.1%)
Chronic kidney disease	20 (11.4%)
Chronic heart disease	12 (6.8%)
Solid organ malignancy	8 (4.5%)
Hematological malignancy	2 (1.1%)
Chronic liver disease	2 (1.1%)
Potential risk factors, n (%)
Exposure to contaminated food/water	14 (8.0%)
Recent travel history	12 (6.8%)
Broad-spectrum antibiotics use	4 (2.3%)
Laboratory findings, median (IQR)
CRP (mg/L)	22.1 (7.2–74.5)
Total WBC count (×103/µl)	7.9 (5.3–12.8)
Relative neutrophil count (%)	73.9 (59.1–86.4)
Relative lymphocyte count (%)	16.5 (6.6–26.2)

Data are presented as numbers (%) for categorical variables and as median (IQR) for continuous variables. IQR, interquartile range; CRP, C-reactive protein; WBC, white blood cell count

Spectrum of enteric pathogens identified in the study cohort

Table 3 summarizes the distribution of detected enteric pathogens. A total of 302 pathogens were identified, including 248 bacterial, 36 viral, and 18 parasitic detections. Overall, bacterial pathogens were identified in 152 patients (86.4%), reflecting a high frequency of bacterial co-detections. Among bacterial pathogens, the most frequently detected were EPEC in 58 patients (33%), EAEC in 40 patients (22.7%), Salmonella spp. in 32 patients (18.2%), enterotoxigenic E. coli (ETEC) in 28 patients (15.9%), and Cl. difficile in 26 patients (14.8%). Viral pathogens were found in 36 patients (20.5%), with norovirus GI/GII was the most common viral agent, identified in 26 patients (14.8%). Parasitic detections were observed in 18 patients (10.2%), predominantly Cryptosporidium spp. (9.1%).

Table 3.

Distribution of enteric pathogens detected by the FilmArray GIP

Pathogens	n (%) of patients (n = 176)
Bacteria, n (%)
EPEC	58 (33.0)
EAEC	40 (22.7)
Salmonella species	32 (18.2)
ETEC	28 (15.9)
Cl. difficile	26 (14.8)
STEC	22 (12.5)
Campylobacter species	20 (11.4)
Shigella/EIEC	12 (6.8)
Plesiomonas shigelloides	8 (4.5)
E. coli O157	2 (1.1)
Viruses, n (%)
Norovirus GI/GII	26 (14.8)
Sapovirus	4 (2.3)
Rotavirus A	4 (2.3)
Adenovirus F40/41	2 (1.1)
Parasites, n (%)
Cryptosporidium species	16 (9.1)
Giardia lamblia	2 (1.1)
Mixed, n (%)
Bacterial	52 (29.5)
Bacterial and viral	24 (13.6)
Bacterial and parasitic	6 (3.4)

Data are presented as numbers (%). Percentages were calculated using the total study cohort (n = 176) as the denominator. EPEC, enteropathogenic Escherichia coli; EAEC, enteroaggregative E. coli; ETEC, enterotoxigenic E. coli; STEC, Shiga toxin-producing E. coli; Cl. difficile, Clostridioides difficile; Shigella/EIEC, Shigella/enteroinvasive E. coli

Single enteric detections were noted in 94 patients (53.4%). In contrast, mixed detections occurred in 82 patients (46.6%), and exhibited substantial heterogeneity, including multiple bacterial pathogens (29.5%), bacterial–viral co-detections (13.6%), and bacterial–parasitic co-detections (3.4%). The most frequent combination was Cl. difficile with EPEC (8 samples). Several samples harbored complex polymicrobial profiles involving three or more pathogens, most commonly diarrheagenic E. coli pathotypes in conjunction with invasive bacteria (e.g., Salmonella spp., Shigella/enteroinvasive E. coli), viruses (norovirus GI/GII), or parasites such as Cryptosporidium spp. and G. lamblia (Fig. 1).

Fig. 1.

Distribution of co-detected enteric pathogens by the FilmArray GIP. Values represent the number of positive stool samples. Cl. difficile, Clostridioides difficile; EPEC, enteropathogenic Escherichia coli; STEC, Shiga toxin-producing E. coli; EAEC, enteroaggregative E. coli; ETEC, enterotoxigenic E. coli; Shigella/EIEC, Shigella/enteroinvasive E. coli. GIP, Gastrointestinal Panel

Seasonal distribution of enteric pathogens detected by the FilmArray GIP

The detected enteric pathogens exhibited temporal variation throughout the study period (Fig. 2). Bacterial agents were identified consistently across most months, with increased frequency during late spring and early autumn, peaking in May (n = 58, 19.2%) and August (n = 28, 9.3%). Viral pathogens showed less pronounced year-round fluctuations, with a relatively higher burden in May (n = 8, 2.6%) and July (n = 7, 2.3%). Parasitic detections remained sporadic, with slight elevations noted in February and August (n = 4, 1.3% each).

Fig. 2.

Month-wise distribution of detected enteric pathogens during the study period. Percentages were calculated using the total number of detected enteric pathogens (n = 302) as the denominator and represent the proportion of each pathogen category per month

Clinical profiles stratified according to detected enteric pathogens by the FilmArray GIP

Enteric pathogens identified by the FilmArray GIP were generally associated with diarrhea, confirming it as the predominant presenting symptom. Watery diarrhea was prevalent among all patients with G. lamblia detection. By contrast, bloody diarrhea was mainly observed in patients with bacterial detections, including Plesiomonas shigelloides (25%), EPEC (10.3%), and Campylobacter spp. (10%). Vomiting was universally reported in all patients with rotavirus A and sapovirus detections, as well as those with selected bacterial pathogens, notably E. coli O157 and P. shigelloides. Fever was present in all adenovirus F40/41 detections and was also frequent among patients with invasive bacterial pathogens, particularly Salmonella spp. (75%), and Campylobacter spp. (70%). Abdominal pain was reported at high rates across most bacterial, viral, and parasitic detections (Fig. 3).

Fig. 3.

Heatmap depicting the distribution of clinical symptoms in patients with single enteric pathogen detection. Values represent the percentage of patients with each detected enteric pathogen who presented with the corresponding clinical symptom. Only patients with single-pathogen detection were included in this analysis (n = 94). Cl. difficile, Clostridioides difficile; EPEC, enteropathogenic Escherichia coli; STEC, Shiga toxin-producing E. coli; EAEC, enteroaggregative E. coli; ETEC, enterotoxigenic E. coli; P. shigelloides, Plesiomonas shigelloides; Shigella/EIEC, Shigella/enteroinvasive E. coli; spp, species

A heatmap of mixed enteric detections further illustrates the complexity of polymicrobial cases (Fig. 4). Patients with bacterial–viral co-detections presented with higher frequencies of vomiting (83.3%), abdominal pain (83.3%), and fever (50%). In contrast, bacterial–parasitic co-detections were associated with the highest frequency of abdominal pain (100%) and watery diarrhea (66.7%). Bloody diarrhea primarily occurred in patients with mixed bacterial detections (7.7%).

Fig. 4.

Heatmap illustrating the distribution of clinical symptoms in patients with mixed enteric pathogen detections. Values represent the percentage of patients with mixed enteric pathogens who presented with the corresponding clinical symptom. Only patients with mixed-pathogen detections were included in this analysis (n = 82)

Agreement between the FilmArray GIP and routine laboratory methods

Table 4 shows the agreement between the FilmArray GIP and stool microscopy for detection of enteric parasites. Stool microscopy failed to identify any parasitic agents, while the multiplex panel detected parasitic pathogens in 18 samples (p < 0.001; Fig. 5). Overall agreement between both techniques was 89.8%, reflecting comparable negative results in most cases. However, agreement for positive detections could not be calculated due to the absence of positive cases by stool microscopy.

Table 4.

Agreement between the FilmArray GIP and stool microscopy for detection of parasites

	Stool microscopy result (n)		Total
	Negative	Positive
FilmArray GIP result (n)
Negative	158	0	158
Positive	18	0	18
Total	176	0	176

Values represent the number of stool samples tested. Agreement refers to the proportion of concordant results between the FilmArray GIP and stool microscopy. GIP, Gastrointestinal Panel

Fig. 5.

Comparison of pathogen detection between the FilmArray GIP and conventional laboratory methods. Values represent the number of stool samples. Statistical significance was assessed using Fisher’s exact test. *p < 0.001, statistically significant. GIP, Gastrointestinal Panel

The FilmArray GIP demonstrated a higher detection yield for Salmonella spp. compared with conventional stool culture (32 versus 6 positive samples), representing a more than five-fold increase (p < 0.001; Fig. 5). Comparison of both methods showed an overall agreement of 85.2%, reflecting concordant negative results (Table 5).

Table 5.

Agreement between the FilmArray GIP and stool culture for detection of Salmonella

	Stool culture result (n)		Total
	Negative	Positive
FilmArray GIP result (n)
Negative	144	0	144
Positive	26	6	32
Total	170	6	176

Values represent the number of stool samples tested. Agreement represents the proportion of concordant results between the FilmArray GIP and stool culture. GIP, Gastrointestinal Panel

Antimicrobial therapy management following the FilmArray GIP results

In this cohort, the FilmArray GIP provided microbiological information that may have supported antimicrobial management decisions; however, therapeutic modifications remained at the discretion of the treating clinicians. Following availability of the FilmArray GIP results, no antimicrobials were prescribed in 53 patients (30.1%). Targeted antimicrobial treatment was newly initiated in 48 cases (27.3%), while continuation of pre-test antimicrobials and escalation were noted in 25 (14.2%) and 19 (10.8%) cases, respectively. Antimicrobial modifications, such as discontinuation with pathogen-directed therapy and de-escalation, were observed in 16 (9.1%) and 10 (5.7%) patients, respectively. Additionally, complete discontinuation of antimicrobials was recorded in 5 (2.8%) cases (Fig. 6).

Fig. 6.

Antimicrobial therapy modifications following the FilmArray GIP testing. Data are presented as percentages of the total cohort (n = 176). GIP, Gastrointestinal Panel

Patterns of empiric and targeted antimicrobial use in the study participants

Of the 176 patients included in this analysis, 75 (42.6%) received empiric therapy and 118 (67%) received targeted antimicrobial treatment, with overlap between the two groups, as some patients received both during their clinical course. In general, ciprofloxacin was the most frequently administered empiric antibiotic, accounting for 25% of prescriptions and also constituted the largest fraction of targeted therapy (23.9%). Ceftriaxone and metronidazole were mainly used empirically (11.4% and 10.2%, respectively), with smaller contributions to targeted regimens. Piperacillin/tazobactam and cefuroxime were prescribed only empirically, though at lower rates (3.4% and 1.7%, respectively). Oral vancomycin was only used as targeted therapy (14.8%), while prescription of other antimicrobials was relatively infrequent (Fig. 7).

Fig. 7.

Distribution of empiric versus targeted antimicrobial use in the study cohort. Data are presented as percentages of empiric and targeted antimicrobial use in the study cohort (n = 176 patients). CIP, ciprofloxacin; CRO, ceftriaxone; MTZ, metronidazole; TZP, piperacillin/tazobactam; CXM, cefuroxime; VA, vancomycin; FEP, cefepime; NTZ, nitazoxanide; AZM, azithromycin; RFX, rifaximin; CFM, cefixime

Specific antimicrobial interventions following the FilmArray GIP results

The most frequent antimicrobial modification associated with the FilmArray GIP findings was continuation of empiric ciprofloxacin monotherapy in 12.5% of patients. Notably, a significant proportion of modifications involved initiation of targeted therapy (25%), most commonly oral vancomycin for Cl. difficile detection (10.2%), rifaximin for travelers’ diarrhea caused by ETEC or EAEC (4.5%), and azithromycin for Campylobacter spp. (4%). Escalation to broader-spectrum antibiotics was recorded in 10.2% of cases, primarily from ceftriaxone to cefepime (8.5%) and from cefuroxime to ciprofloxacin (1.7%). These modifications were guided by the overall clinical context, including disease severity and patient-related factors, alongside the FilmArray GIP findings, as the assay does not provide antimicrobial susceptibility data. Empiric therapy was discontinued in favor of pathogen identification, with 1.7% of patients each switched from piperacillin/tazobactam or ciprofloxacin to vancomycin. In addition, de-escalation was observed among our cohort, as reflected by cessation of metronidazole in combination therapy (2.3%) and substitution of piperacillin/tazobactam with ciprofloxacin in 1.7% of cases (Table 6).

Table 6.

Most frequent antimicrobial modifications following the FilmArray GIP testing

Antimicrobial stewardship action	Empiric therapy	Post-BioFire therapy	n (%)
Continued	Ciprofloxacin	Ciprofloxacin	22 (12.5)
Targeted initiation	–	Vancomycin	18 (10.2)
	–	Rifaximin	8 (4.5)
	–	Azithromycin	7 (4.0)
	–	Ciprofloxacin	5 (2.8)
	–	Nitazoxanide	5 (2.8)
	–	Cefixime	3 (1.7)
Escalation	Ceftriaxone	Cefepime	15 (8.5)
	Cefuroxime	Ciprofloxacin	3 (1.7)
Discontinuation with pathogen-directed therapy	Ciprofloxacin + Metronidazole	Nitazoxanide	5 (2.8)
	Ciprofloxacin	Vancomycin	3 (1.7)
	Piperacillin/tazobactam	Vancomycin	3 (1.7)
	Ciprofloxacin	Azithromycin	2 (1.1)
De-escalation	Ciprofloxacin + Metronidazole	Ciprofloxacin	4 (2.3)
	Piperacillin/tazobactam	Ciprofloxacin	3 (1.7)

Data are presented as numbers (%) of patients in the study cohort (n = 176). GIP, Gastrointestinal Panel

Transmission-based isolation according to pathogens detected by the FilmArray GIP

Among 176 patients with positive FilmArray GIP results, 15 (8.5%) were subjected to transmission-based isolation precautions. Isolation was implemented in 11 patients with Cl. difficile (42.3%) versus 4 with norovirus detections (15.4%), with no statistically significant difference between both groups (p = 0.11). Given the relatively small number of isolated cases, this comparison should be interpreted cautiously, as the study may have been underpowered to detect statistically significant differences. It is worth noting that transmission-based isolation was not initiated in patients with Cl. difficile–norovirus co-detections or those with other detected enteric pathogens (Fig. 8).

Fig. 8.

Transmission-based isolation status following the FilmArray GIP. Values represent the number of patients. Only cases with Cl. difficile or norovirus detections underwent transmission-based isolation. The “combined” category includes cases with Cl. difficile-norovirus co-detections. “Others” represents cases with all remaining positive samples, which may involve single or multiple pathogen detections. Cl. difficile, Clostridioides difficile. GIP, Gastrointestinal Panel

Clinical course and outcomes of patients with single and mixed enteric detections

The clinical course of AGE was comparable between patients with single and mixed enteric pathogens. In particular, rates of hospital (36.2% versus 31.7%) and ICU admission (10.6% versus 12.2%) were similar between the two groups. The frequency of complications, including volume depletion and septic shock, also did not differ significantly across groups (p = 0.95 and 0.84, respectively). Rates of clinical improvement were high in both groups, with all patients in the single-detection group recovering (100%) and a similarly favorable outcome was observed in the mixed-detection cohort (97.6%). Overall mortality was low and limited to two patients (2.4%) in the mixed-detection group (Table 7).

Table 7.

Clinical course and outcomes of patients with single versus mixed enteric detections

Variables	Single detection (n = 94)	Mixed detection (n = 82)	OR (95% CI)	p-value
Hospital admission, n (%)	34 (36.2%)	26 (31.7%)	0.82 (0.44–1.53)	0.53
LOS (days), median (IQR)	0.0 (0.0–3.0)	0.0 (0.0–2.25)	NA	0.47
ICU admission, n (%)	10 (10.6%)	10 (12.2%)	1.16 (0.46–2.96)	0.74
Complications, n (%)
Volume depletion	34 (36.2%)	30 (36.6%)	1.02 (0.55–1.88)	0.95
Septic shock	4 (4.3%)	4 (4.9%)	1.15 (0.27–4.76)	0.84
Clinical outcomes, n (%)
Recovered	94 (100.0%)	80 (97.6%)	–	0.22
Deceased	0 (0.0%)	2 (2.4%)

Data are presented as numbers (%) for categorical variables and as median (IQR) for continuous variables. Categorical variables were compared using the chi-squared (χ²) test, while continuous variables were compared using the Mann–Whitney U test. ICU length of stay was not compared between groups due to a zero-inflated distribution (most patients had no ICU admission). Odds ratios were not assessed for outcomes with zero events in one group. LOS, length of hospital stay; IQR, interquartile range; OR, odds ratio; CI, confidence interval; NA, not applicable

Discussion

The present study provides further evidence supporting the role of syndromic molecular testing in clinical management of adult AGE. In 176 patients (mean age 44.9 ± 19.7 years), bacterial pathogens were the most frequently detected (86.4%), followed by viral agents (20.5%) and parasites (10.2%). This microbial pattern mirrors previous studies, confirming the predominance of bacterial etiologies in adult acute diarrhea, followed by viral and parasitic causes [14–15].

Among bacterial agents, diarrheagenic E. coli pathotypes (EPEC, EAEC, and ETEC) were the most prevalent. This finding aligns with another Saudi Arabian study [8] and underlines the utility of molecular platforms in detecting pathogens that may be overlooked by conventional techniques. By contrast, Cl. difficile (40.9%) and Campylobacter spp. (15.9%) were the most frequent pathogens in a Korean study. This discrepancy may be attributed to different inclusion criteria (adult-only patients in our study versus all age groups), and contribution of hospitalized patients to the majority of samples, a population at increased risk of Cl. difficile infection [16]. As a nucleic acid amplification test, the FilmArray GIP can detect bacterial DNA regardless of organism viability. Therefore, pathogen detection does not necessarily reflect active infection or establish causality. Several enteric targets detected by multiplex PCR, such as EPEC, EAEC, and Cl. difficile may represent incidental carriage or colonization, particularly in adults, as reported recently [17]. Accordingly, correlation with clinical findings and disease severity is crucial to reduce overdiagnosis and potential treatment of colonization. Of note, other confirmatory molecular techniques and Cl. difficile toxin assay were not available in our workflow, and multiplex molecular testing results were interpreted in conjunction with compatible clinical symptoms documented in the medical records.

Norovirus GI/GII predominated among viral etiologies in our cohort (14.8%), consistent with previous studies (3.8%) [18]. Interestingly, 15.4% of patients with norovirus detection (4/26) reported recent travel, in line with findings among travelers in the United States and Europe [19]. In addition, less frequently reported viral agents, including sapovirus and rotavirus A were detected in our cohort (each 11.1%). Although often under-detected in adult populations, these viruses have emerged as causes of enteric infections with the implementation of multiplex molecular panels, as observed recently [7]. The ability of the FilmArray GIP to detect less commonly encountered viral pathogens underscores its value in providing better understanding of the epidemiology of enteric pathogens. In this cohort, parasitic pathogens accounted for 10.2% of detections, with Cryptosporidium representing the majority, concordant with a previous study [20]. The increased detection of this parasite highlights the importance of definite pathogen identification to ensure timely initiation of antiparasitic therapy, particularly in high-risk groups.

Mixed enteric pathogens were identified in a substantial proportion of the cohort (46.6%). These polymicrobial detections frequently involved multiple E. coli types, either alone or in combination with other bacteria, viruses, or parasites. Interestingly, two stool samples uniquely harbored up to five pathogens (EPEC, EAEC, ETEC, Shigella/EIEC, and norovirus GI/GII). Such complexity is unlikely to be captured by conventional diagnostic approaches or single-target assays. Our findings are consistent with a multicenter study from Lebanon, whereby mixed enteric detections accounted for 54% of cases [21]. Another recent observational study in high-risk adults reported that co-detections most frequently involved EPEC alongside other enteric organisms, suggesting a frequent association of diarrheagenic E. coli with other enteric pathogens [22]. It is noteworthy that multiplex molecular co-detections do not necessarily equate to clinical infection. Their significance remains uncertain and may reflect colonization or asymptomatic shedding, depending on the clinical context. Accordingly, these results should be interpreted in light of this limitation.

Seasonal dynamics in the pattern of enteric pathogens were noted during the study period, with both bacterial and viral detections recorded throughout the year. These findings underscore the role of syndromic molecular testing in providing comprehensive year-round microbiological surveillance. Relatively higher rates of bacterial detections were observed during warmer months. Similarly, previous studies from hot-climate regions have described an increased burden of enteric bacterial pathogens during warmer periods [23]. However, since inferential statistical analyses (e.g., time-series or regression models) evaluating climatic associations were not performed in this study, these findings should be interpreted cautiously and not considered evidence of a direct causal relationship.

Given the marked overlap in the clinical manifestations of AGE shown in our cohort, reliance exclusively on symptoms is insufficient to predict the causative organisms. Therefore, the ability of multiplex platforms to bridge clinical presentation with accurate etiologic diagnosis supports more targeted therapeutic decisions, with potential contribution to antimicrobial stewardship. This finding aligns with previous studies showing associations between specific FilmArray GIP targets and clinical symptoms. For example, STEC and Campylobacter spp. have been more commonly detected in patients presenting with hemorrhagic diarrhea, suggesting a possible pathogen–clinical correlation that may be further explored using molecular testing [24].

While stool microscopy failed to identify any parasites, the multiplex platform detected parasitic targets in 18 samples (16 Cryptosporidium spp. and 2 G. lamblia), with an overall agreement of 89.8%. Our observations are consistent with other studies indicating that routine microscopy is prone to false-negative results, especially in patients with low parasitic load, intermittent shedding, and poor sample quality [25]. In this context, a recent study conducted among critically ill patients in Egypt reported a considerably improved detection of Cryptosporidium and G. lamblia using the FilmArray GIP [26]. Similarly, a meta-analysis has demonstrated high diagnostic performance of multiplex platforms for most evaluated enteric pathogens compared with conventional methodologies [27]. Overall, these findings highlight the higher detection yield of multiplex molecular platforms compared with routine microscopy, while acknowledging the limitations of stool microscopy as a reference method for detection of enteric parasites.

The FilmArray GIP testing showed a more than five-fold increase in detection of Salmonella spp. compared with stool culture (18.2% versus 3.4%; p < 0.001). The lower yield observed with conventional stool culture may reflect its limited diagnostic performance, particularly in cases with low bacterial burden, prior antibiotic exposure, or suboptimal sample processing [28]. Similarly, a recent large-scale study across 13 sites in Europe and the United States reported improved detection of Salmonella spp. using multiplex platforms compared with traditional culture-based testing [29]. In our study, the overall 85.2% agreement between the FilmArray GIP and stool culture should be interpreted cautiously, as conventional culture techniques represent imperfect comparator, particularly in the absence of additional confirmatory methods such as alternative PCR assays or genomic sequencing. Importantly, other bacterial pathogens detected by the panel are not routinely cultured in our microbiology laboratory; therefore, direct comparison with conventional methods was restricted to Salmonella spp.

In this cohort, the FilmArray GIP testing was potentially associated with modifications in antimicrobial prescribing, with 30.1% of patients avoiding unnecessary antibiotics and 27.3% initiating targeted therapy. These results are in line with the growing body of evidence that syndromic panels contribute to optimized antimicrobial management. For example, a recent retrospective analysis demonstrated that the use of multiplex PCR increased definitive antimicrobial use in 66.7% of the post-panel group [30]. Similarly, a study conducted at an HIV Reference Center in Brazil reported a 30% reduction in antibiotic use following the FilmArray GIP testing, with a significant reduction in antimicrobial days [31]. Together, these findings indicate that integrating rapid molecular panels into standard diagnostics provides a pragmatic approach toward rational antimicrobial use.

The total proportion of patients who either initiated targeted oral vancomycin (10.2%) or had ciprofloxacin discontinued in favor of vancomycin (1.7%) closely corresponds with the prevalence of identified Cl. difficile detections (14.8%). These findings reflect efforts toward appropriate antimicrobial use associated with the FilmArray GIP results. Comparable treatment modifications have also been described in the literature [32]. The discrepancy between molecular detection and vancomycin prescribing likely suggests personalized clinical decision-making, possibly related to self-limited disease, colonization rather than true infection, or contraindications to vancomycin.

Rifaximin, characterized by poor systemic absorption and increased intraluminal concentration, is commonly used for travelers’ diarrhea caused by ETEC or EAEC [33]. In this study, rifaximin was initiated in 4.5% of patients following the FilmArray GIP results, emphasizing its role as a gut-selective antibiotic. Given the growing prevalence of fluoroquinolone resistance, the Infectious Diseases Society of America (IDSA) recommends a five-day azithromycin regimen for treatment of Campylobacter infections [34]. In this cohort, empiric ciprofloxacin was switched to azithromycin in 1.1% of patients following molecular testing, whereas the latter was initiated de novo in 4%. Comparable modifications have been reported in a Romanian study, although with different prescribing rates [35]. The observed escalation and de-escalation patterns further suggest tailored antimicrobial efforts following the FilmArray GIP results. Similarly, a recent study demonstrated that rapid molecular testing was associated with antimicrobial changes in almost one‑third of patients [36].

The FilmArray GIP findings were temporally associated with infection control-related decisions, with 8.5% of patients undergoing transmission-based isolation precautions. Patients with Cl. difficile detections were placed under isolation at a higher rate than those with norovirus (42.3% versus 15.4%; p = 0.11). The high transmissibility and spore-forming ability of this organism are likely causes of different isolation rates. Similarly, a previous study reported that traditional diagnostic techniques failed to detect pathogens that warrant isolation, while the multiplex panel enhanced pathogen detection and supported isolation decisions in adult patients [37]. Additionally, a study from a Dutch tertiary care center underlined the role of multiplex panels in facilitating earlier termination or refinement of isolation precautions relative to routine PCR testing [38]. It is noteworthy that isolation decisions among our patients were also determined by clinical assessment and hospital infection control guidelines rather than microbiological detection alone.

Interestingly, the detection of mixed enteric pathogens did not translate into worse clinical outcomes. In our study, patients with single versus multiple enteric pathogens exhibited comparable rates of hospital and ICU admission, complications, and overall recovery. These findings support emerging evidence that co-detections via multiplex panels are not invariably associated with greater disease severity [39]. The favorable outcomes and low mortality observed in both groups imply that clinical severity and management are driven more by the clinically relevant organism and overall clinical context rather than by the presence of multiple enteric pathogens. However, these conclusions should be interpreted with caution given the relatively small sample size and low event rates, which may have limited the ability to detect clinically meaningful differences between groups.

Study limitations

Our study has several limitations that warrant consideration. The retrospective design may have resulted in selection and information biases, as well as limited causal inference. In addition, inclusion was limited to patients tested with the FilmArray GIP, which may introduce physician selection bias, as testing was more likely conducted in more severe or atypical cases. Accordingly, the findings may not be fully generalizable to all patients with AGE. The lack of a control group (e.g., pre-implementation or non-tested cohort) further restricted comparative assessment of the findings. As a single-center study, the findings may not fully represent the broader Saudi population, potentially limiting generalizability. Additionally, pathogen detection via the FilmArray GIP did not necessarily reflect active infection, as molecular assays may identify asymptomatic carriage or prolonged shedding, especially in mixed detections. Another limitation is the absence of confirmatory molecular or toxin assays that may have limited validation of the FilmArray GIP findings. Furthermore, the FilmArray GIP does not provide antimicrobial susceptibility data, which may limit direct therapeutic guidance. Finally, incomplete documentation of antimicrobial duration limited comprehensive evaluation of treatment outcomes.

Conclusion

In our cohort, implementation of the FilmArray GIP contributed to increased detection of enteric pathogens compared with conventional diagnostic workflows. In addition, its use was associated with pathogen-directed antimicrobial modifications and initiation of infection control practices within the clinical setting. Overall, the FilmArray GIP may serve as a useful adjunct diagnostic modality, particularly in well-resourced settings, supporting timely pathogen detection and clinical decision-making.

Abbreviations

AGE

Acute gastroenteritis

Cl.difficile

Clostridioides difficile

EAEC

Enteroaggregative Escherichia coli

EIEC

Enteroinvasive Escherichia coli

EPEC

Enteropathogenic Escherichia coli

ETEC

Enterotoxigenic Escherichia coli

FilmArray GIP

FilmArray Gastrointestinal Panel

G. lamblia

Giardia lamblia

STEC

Shiga toxin-producing Escherichia coli

Author contributions

RAE: Project administration, Conceptualization, Methodology, Writing - review & editing, Supervision, Investigation, Resources. AA: Writing - original draft, Formal analysis, Data curation. RME: Data curation, Writing - original draft, Formal analysis. All authors have read and approved the final version of the manuscript.

Funding

The authors received no financial support for this research.

Data availability

All data generated or analyzed during this study are included in this published article.

Declarations

Ethics approval and consent to participate

The study followed the ethical standards of the institutional and national research committee and the tenets of the 1964 Declaration of Helsinki and its later amendments. The study protocol was reviewed and approved by the Institutional Review Board (IRB) of Fakeeh Care Group (reference number FIRB-25-0033). All data were anonymized before analysis, and patient confidentiality was strictly maintained throughout the study. The requirement for informed consent was waived due to the retrospective nature of the study and the use of de-identified dataset.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.