Confirmation of the POLA Index and Its Association with COVID-19 and Influenza Incidence, Gut Microbiota, and Mucosal Immunity Markers in Young Women: A Pilot Cohort Study

Abstract

Evidence for diet-driven immunomodulation remains limited in terms of practical tools. This pilot study prospectively evaluated the dietary POLA index in 51 women aged 25–45 years and living in Kraków, Poland. At baseline (2022), 7-day dietary intake and physical activity were recorded; stool samples were analyzed for gut microbiota composition; short-chain fatty acids were quantified using high-performance liquid chromatography; and fecal secretory IgA (sIgA), human β-defensin-2 (HBD-2), and calprotectin levels were measured by the enzyme-linked immuno-sorbent assay. In June 2023, post-baseline incidence of COVID-19 and influenza was self-reported. Participants were categorized into a group with beneficial immunomodulation (BIM, POLA score ≤ 5) and a group with unbeneficial/highly unbeneficial immunomodulation (UBIM + HUBIM, POLA score > 5). The incidence of COVID-19 or influenza was 7.7% (1/13) in the BIM group vs. 36.8% (14/38) in the UBIM + HUBIM group. After adjusting for age and smoking, the UBIM + HUBIM group had higher odds of infection compared with the BIM group (adjusted OR = 6.53; 95% CI: 1.02–129.85), corresponding to a higher absolute risk of 47.9% (95% CI 26.1–70.6) versus 12.3% (95% CI 1.4–58.1). The BIM group more often met the fiber and micronutrient adequacy targets and showed a higher proportion of sIgA levels within the reference range (92.3% vs. 60.5%), along with lower fecal succinic acid concentrations (median 3.27 mg/g vs. 4.32 mg/g). In this cohort, a favorable POLA score was associated with lower self-reported COVID-19 or influenza incidence and an sIgA profile suggestive of intestinal immune homeostasis. As this pilot study is underpowered, findings should be interpreted as exploratory and hypothesis-generating. Nonetheless, results support the POLA index as a practical diet-quality metric with potential immunomodulatory relevance.

1. Introduction

Coronavirus disease 2019 (COVID-19) is an infectious illness caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). It was officially declared a pandemic in March 2020 and has since profoundly affected populations worldwide. Over the past four years, the cumulative number of infections has exceeded 775 million [1]. The severity of COVID-19 has been a major determinant of virus-related mortality. Recent data indicate that the infection fatality rate (IFR) varies substantially across populations, age groups, and geographic locations, with an average global IFR of 0.15%, and a slightly lower IFR of 0.095% for individuals aged 0–69 years [2,3]. In contrast, among hospitalized patients, the IFR increases to 13.2% and it rises sharply to 55.9% in those requiring mechanical ventilation [4].

Influenza is an acute respiratory illness caused by influenza A or B viruses, typically presenting with symptoms similar to those of the common cold. In clinical practice, however, the term is often used imprecisely to describe a constellation of nonspecific inflammatory symptoms resulting from infections with various upper respiratory viruses or even non-infectious causes. The diagnostic accuracy of self-administered polymerase chain reaction testing is limited, with reported sensitivities as low as 58% among symptomatic individuals [5]. Moreover, in nearly 60% of patients presenting with influenza-like symptoms, no causative pathogen can be identified [6]. Neither individual clinical signs, nor their combinations, nor current case definitions of influenza-like illness provide sufficient accuracy for reliably diagnosing true influenza infections [7]. Consequently, the term “influenza” is often used to denote a general inflammatory state rather than a specific viral etiology. It has been suggested that viral infections follow two primary pathogenic pathways. The first involves the inoculation and replication of the virus in the upper respiratory tract, which may or may not extend to the lungs. The second entails a host immune response characterized by an extrapulmonary systemic hyperinflammatory syndrome [8].

To date, aside from antibiotic use and ventilator support, only a limited number of therapies have demonstrated modest clinical benefits for patients with COVID-19. Evidence suggests that various nonpharmacological interventions, including physical activity, nutrition, psychological support, and lifestyle modifications, can serve as effective strategies for managing inflammation [9].

The Mediterranean diet is widely recognized for its anti-inflammatory properties. It emphasizes a high intake of fruits, vegetables, olive oil, whole grains, and nuts, along with moderate to low consumption of fish, poultry, fermented dairy products, wine, and red or processed meats [10]. Adherence to this dietary pattern provides a wide range of anti-inflammatory and immunomodulatory nutrients, including essential vitamins (A, B, C, D, and E) and minerals such as copper, selenium, and zinc, all of which support nutritional well-being [11]. Available evidence suggests that greater adherence to a balanced, anti-inflammatory diet rich in vegetables, fruits, and nuts, such as the Mediterranean diet may be linked to a reduced risk of SARS-CoV-2 infection and severe illness [12]. A study of 23 countries reported an inverse association between adherence to the Mediterranean diet and SARS-CoV-2-related mortality, even after adjusting for confounding factors such as physical inactivity and well-being [13].

Researchers have proposed a potential link between diet and susceptibility to viral infections. Vegetables, fruits and fish oil—key components of the Mediterranean diet—are rich sources of vitamins A, C, and D. These vitamins act as antioxidants against reactive oxygen species produced by immune cells, thereby helping to preserve cellular integrity and maintain epithelial barrier function. Vitamin D, in particular, stimulates polymorphonuclear and natural killer cells to produce potent antimicrobial peptides. These immune effects contribute to a reduced risk of infection and enhanced viral clearance. As a result, there is growing support for the protective role of the Mediterranean diet in the context of COVID-19 [14,15].

Although the concept of immunomodulation is well established, practical tools assessing the impact of current dietary patterns on immune function are lacking. One of the widely recognized tools is the Dietary Inflammation Index (DII), which evaluates the association between dietary intake and inflammation levels [16]. In 2022, a novel metric—the POLA index—was developed [17]. While the DII primarily focuses on the anti-inflammatory effects of diet, the POLA index also incorporates the influence of various bioactive compounds found predominantly in vegetables, fruits, and nuts—compounds for which most nutritional databases lack comprehensive data. These bioactives exert multiple immunomodulatory effects, including anti-inflammatory, antiviral, and antibacterial actions, and promote the growth of beneficial gut microbiota [15].

The POLA index shows promise as a tool for mitigating risks associated with SARS-CoV-2 infection. Therefore, the aim of this study was to validate the POLA index in a new cohort of individuals. In addition, we sought to examine its associations with short-chain fatty acids (SCFAs) and fecal markers of intestinal inflammation and mucosal immunity.

2. Materials and Methods

2.1. Study Design and Participants

This prospective observational study was conducted between September and December 2022 at the Department of Nutrition and Drug Research, Jagiellonian University Medical College (Kraków, Poland), as part of an institutional research program. We enrolled healthy women (without comorbidities) aged 25–45 years who had followed either a vegetarian or traditional omnivorous diet for at least 1 year. Recruitment was conducted in Kraków via social media announcements and snowball sampling. Interested individuals contacted the study team and were screened for eligibility. To reflect habitual lifestyle behaviors, each participant underwent a 1-week observation period. If a special event (e.g., a wedding or birthday) was scheduled, observation began the following week. During the observation period, participants were instructed to maintain their usual diet and physical activity patterns. Inclusion criteria were as follows: age 25–45 years; absence of chronic diseases; body mass index (BMI) between 18.5 and 29.9 kg/m²; adherence to a stable vegetarian or traditional omnivorous diet for at least 12 months; and no use of antibiotics or probiotics in the 3 months preceding enrollment. A total of 51 women were included (traditional diet: n = 36; vegetarian diet: n = 15). All participants provided written informed consent to take part in the study.

2.2. Anthropometry and Body Composition

To minimize measurement error, participants were instructed to empty their bladder if needed and to avoid vigorous activity for 24 h prior to testing. Stature was measured to the nearest 0.1 cm using a Seca 213 portable stadiometer under standardized conditions (barefoot, standing upright). Body weight and segmental body composition were assessed using a calibrated bioelectrical impedance analyzer (Tanita BC-418 MA, Tokyo, Japan; 6.25 and 50 kHz; 90 µA), with measurements recorded to the nearest 0.1 kg and 0.1%, respectively. BMI was calculated as weight (kg) divided by height in squared (m²).

2.3. Physical Activity Monitoring

Habitual physical activity and sleep were monitored over 7 consecutive days using a Polar M430 wrist-worn device, worn on the nondominant hand. Participants were instructed to wear the device continuously (24 h/day), removing it only during bathing. The device provided data on total energy expenditure, step count, and sleep duration. Both written and verbal instructions for device use were provided to each participant.

2.4. Dietary Assessment

Prior to the observation week, participants received detailed instructions and examples on how to maintain a 7-day prospective food diary. They were encouraged to use kitchen scales or a validated photographic portion guide to enhance the accuracy of portion size estimation. All foods and beverages consumed were recorded, and the diaries were reviewed with participants at the end of the week to ensure completeness. Dietary records were entered into the “Dieta 6.0” software (National Food and Nutrition Institute, Warsaw, Poland) to calculate daily energy and nutrient intakes, including macronutrients, fiber, cholesterol, vitamins, and minerals. Nutrient adequacy was assessed against Polish age- and sex-specific reference values: Recommended Dietary Allowance (RDA), Adequate Intake (AI), or Estimated Average Requirement (EAR), as appropriate [18].

2.5. Gut Microbiota (Culture-Based)

Fecal microbiota composition was assessed using both qualitative and quantitative culture-based microbiological methods. Participants collected stool samples at home using sterile fecal tubes and delivered them to the microbiology laboratory within 6 h under refrigerated conditions. Samples were accurately weighed, suspended in Schaedler’s liquid medium (Oxoid Ltd., Basingstoke, UK), and thoroughly homogenized using sterile glass beads. Decimal dilution was then performed to quantify colony-forming units per gram of stool. Diluted samples were plated onto selective, differential, and chromogenic agar media. Incubation was conducted at 37 °C for 24 h for aerobic bacteria and 48–72 h for anaerobic bacteria. Following incubation, microbial colonies were identified using biochemical assays (API, bioMérieux, l’Etoile, France) and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI Biotyper, Bruker Scientific LLC, Billerica, MA, USA) according to the manufacturer’s protocols. The analysis included the following enteric microbes: Escherichia coli, Escherichia coli Biovare, Providencia spp., Pseudomonas spp., Klebsiella spp., Enterobacter spp., Citrobacter spp., Serratia spp., Enterococcus spp., Streptococcus spp., Bacillus spp., Bifidobacterium spp., Bacteroides spp., Lactobacillus spp., H₂O₂-producing Lactobacillus, and Clostridium spp.

2.6. Fecal Markers of Intestinal Inflammation and Immunity

Fecal concentrations of intestinal inflammation and mucosal immunity markers, including secretory immunoglobulin A (sIgA), human beta-defensin 2 (HBD-2), and calprotectin, were assessed. The sIgA level was examined using the IDK^®sIgA ELISA kit (cat. no. K 8870; Immundiagnostik AG, Bensheim, Germany) according to the manufacturer’s protocol. Fecal samples (15 mg) were mixed with 1.5 mL of wash buffer and allowed to stand for approximately 10 min until sediment settled. The supernatant was then diluted to a final concentration of 1:12,500. A 100 µL aliquot of the diluted sample was processed using the sandwich ELISA principle. Microplates were first washed 5 times with 250 µL of wash buffer, and excess liquid was absorbed. Diluted samples, standards, and controls were added to microtiter wells coated with polyclonal rabbit anti-human IgA antibodies and incubated for 1 h on a horizontal shaker at room temperature. After washing, 100 µL of peroxidase-conjugated mouse anti-sIgA antibody was added. Following a second incubation and wash cycle, tetramethylbenzidine substrate was added and incubated for 15 min in the dark. The reaction was then stopped with an acidic stop solution. All samples were immediately analyzed using a Stat Fax 2100 microplate reader (Awareness Technology, Inc., Palm City, FL, USA) at 450 nm with a reference wavelength of 620 nm. Secretory IgA levels were reported in µg/g, with reference values defined as follows: <500 µg/g, subnormal; 500–2040 µg/g, normal; and >2040 µg/g, elevated. HBD-2 levels were measured using a quantitative enzyme immunoassay (cat. no. K 6500 β-defensin 2 ELISA, Immunodiagnostik AG), while calprotectin levels were determined using the Calprotectin ELISA kit (cat. no. EQ 6831-9601 W; Euroimmun Medizinische Labordiagnostika AG, Lübeck, Germany).

2.7. Short-Chain Fatty Acids by High-Performance Liquid Chromatography

After thawing, fecal samples were dried to a constant weight at 40 °C and then homogenized. Short-chain fatty acid extraction was performed following the method described by Dobrowolska-Iwanek et al. [19], with minor modifications. Briefly, 3 portions of 0.1 g each were weighed from every fecal sample. Each portion was mixed with 1 mL of 10% perchloric acid, subjected to ultrasonic treatment for 40 min (ultrasonic bath, 400 W, ULTRON U-503; Ultron, Dywity, Poland), and then shaken for 4 min. The extraction procedure was repeated twice on the remaining stool pellets, and the supernatants were collected after each extraction. The combined supernatants were adjusted to a final volume of 3 mL with 10% perchloric acid and centrifuged for 15 min at 9000 rpm. Prior to analysis, the resulting supernatant was filtered through a 0.45-µm nylon syringe filter (17 mm diameter; La-Pha-Pack, Langerwehe, Germany).

Extracts were diluted 5, 10 or 15 times with 10% perchloric acid so that analyte concentrations fell within the calibration curve. Quantification of organic acids was performed using a standard curve approach. Calibration solutions were prepared at the following concentrations: 1 mg/mL, 0.5 mg/mL, 0.25 mg/mL, 0.125 mg/mL, 0.0625 mg/mL, 0.0312 mg/mL, and 0.0156 mg/mL. The determined R² values were all above 0.998 for each organic acid The lowest concentration (0.0156 mg/mL) was established as the limit of quantification for the assay.

The determination of organic acids was conducted using a high-performance liquid chromatography (HPLC) system, equipped with a PDA 100 UV–VIS detector, a P680 pump, a TCC100 thermostat, and an ASI100 autosampler, controlled by CHROMELEON^® 6.60 software (Dionex; Thermo Fisher Scientific, Bremen, Germany). Chromatographic separation was performed using a Synergi 4 µm Hydro-RP 80A C18 column (250 mm × 4.6 mm, 4-µm particle size) (Phenomenex’s, Torrance, CA, USA)maintained at 35 °C. The HPLC system operated at a flow rate of 1 mL/min in gradient mode, using a mobile phase composed of (A) acetonitrile and (B) 0.2% orthophosphoric acid. The gradient program, based on the method described by Dobrowolska-Iwanek et al. [20], was as follows: 0 min—A: 5%; 6 min—A: 20%; 12 min—A: 30%; 16 min—A: 30%; 25 min—A: 40%; 28 min—A: 80%; 32 min—A: 80%; and 35 min—A: 5%. Each sample was introduced into the system in duplicate injections each with an injection volume of 20 µL. To ensure the stability of the analytical response, a selected standard solution was analyzed following each set of 10 extract measurements. Target analytes included acetic acid, propionic acid, butyric acid, isovaleric acid, and succinic acid. Final concentrations were expressed as mg/g of dry fecal mass.

2.8. POLA Index Calculation

Dietary intake data from the 7-day observation period were used to calculate the POLA index for each participant. Only participants with at least 5 valid recording days (a minimum of 3 weekdays and 2 weekend days) were included in the analysis. Average daily intake across all available days was used. The POLA index incorporates the following dietary components: potassium, magnesium, iron, zinc, calcium, vitamin A, vitamin E, thiamin, vitamin B6, vitamin C, vitamin D, linoleic acid (LA), α-linolenic acid (ALA), folates, fiber, moreover fruit, vegetable and nut consumption. For each component, a score of 0 points was assigned if the participant’s mean intake met or exceeded 100% of the age- and sex-specific reference value (RDA, AI, or EAR), and 1 point if intake was below the reference. Two components were scored using a modified approach: (1) vitamin D intake was categorized into three levels: ≥100%, 50–<100%, or <50% of the reference, scoring 0, 1, or 2 points, respectively; (2) combined fruit and vegetable intake was also categorized into three levels: ≥600 g/day, 400–<600 g/day, or <400 g/day, scoring 0, 1, or 2 points, respectively. This adjustment reflects the importance of whole-food bioactives not captured by standard nutrient databases. Additionally, nut intake was scored as follows: ≥10 g/day—0 points and <10 g/day—1 point. Total POLA scores were calculated by summing all component points for each participant. Scores were categorized as follows: ≤5 points—beneficial immunomodulation (BIM), indicating a diet with a favorable immunomodulatory profile; 6–11 points—unbeneficial immunomodulation (UBIM); and ≥12 points—highly unbeneficial immunomodulation (HUBIM). For statistical analysis, UBIM and HUBIM categories were combined into a single group (UBIM + HUBIM).

2.9. Statistical Analysis

Descriptive statistics were presented as means with standard deviations or medians with interquartile ranges (Q1–Q3), as appropriate. The normality of continuous variables was assessed with the Shapiro–Wilk test. Between-group differences in quantitative or ordinal variables were evaluated using either Student’s t-test or the Mann–Whitney U test. Differences in categorical variables were assessed using the Chi-square test or Fisher’s exact test, as appropriate. Correlations between quantitative variables were examined using Spearman’s rank correlation coefficient. Multivariable logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs). Absolute risks (predicted probabilities) were derived from multivariable logistic regression models by calculating the inverse logit of the linear predictor. To obtain marginal absolute risks for categories of the POLA index, predictions were averaged over the empirical joint distribution of all covariates (population-averaged marginal standardization). All analyses were conducted using PS IMAGO PRO 10 (IBM SPSS Statistics 29), STATISTICA 13.3, and R version 4.0.4. A p-value of less than 0.05 was considered statistically significant.

2.10. Follow-Up Outcome Ascertainment

In June 2023 (approximately 6 months after the observation period), all 51 participants were contacted to determine whether they had developed COVID-19 or influenza since baseline. Participation in this follow-up survey was voluntary, and all questions were optional; however, the response rate was 100%. A total of 15 participants (29.4%) reported having experienced either COVID-19 or influenza during the follow-up period. These self-reported outcomes (no laboratory confirmation or formal case definition was applied) were analyzed in relation to baseline dietary behaviors (including the POLA index category), physical activity, and gut microbiota parameters.

3. Results

3.1. Characteristics of Participants

Participant characteristics stratified by POLA index category are presented in

Table 1. Characteristics of participants according to POLA index category (BIM vs. UBIM + HUBIM).

Variable		BIM n = 13		UBIM + HUBIM n = 38		Student’s t-Test
		$\overline{\mathbf{X}}$ ± SD		$\overline{\mathbf{X}}$ ± SD		p-Value
Age [years]		33.9 ± 5.0		35.4 ± 5.8		0.3886
Body weight [kg]		58.8 ± 5.7		63.3 ± 7.5		0.0578
Height [cm]		165.8 ± 5		166.9 ± 6.2		0.5567
BMI [kg/m2]		21.3 ± 1.8		22.7 ± 2.5		0.0836
Body fat [%]		25.5 ± 4.6		28.7 ± 5.8		0.0539
TEE [kcal]		2086.7 ± 315.8		2098.8 ± 227.3		0.9003
PAL		1.53 ± 0.2		1.50 ± 0.14		0.5748
Sleep duration [h]		7:42 ± 0:30		7:31 ± 0:51		0.3497
Steps		14,765 ± 4554		13,161 ± 3944		0.2717
POLA score [points]		2.92 ± 1.71		10.2 ± 3.0		<0.0001
Variable	Category	n	%	n	%	Chi2 p-value
Diet type	Traditional	8	61.5	28	73.7	0.4068
	Vegetarian	5	38.5	10	26.3
BMI category	Normal body weight	13	100	30	78.9	0.0958
	Overweight	0	0	8	21.1
Marital status	Single/divorced	9	69.2	15	39.5	0.1760
	Married/cohabiting	4	30.8	23	60.5
Body fat category	Underfat	3	23.1	4	10.5	0.0904
	Normal	10	76.9	24	63.2
	Overfat	0	0.0	10	26.3
Smoking status	No	13	100	32	84.2	0.1272
	Yes	0	0	6	15.8
Education level	Secondary	0	0	2	5.3	0.3987
	Higher	13	100	36	94.7
Self-rated physical activity (leisure)	Low	3	23.1	9	23.7	0.1152
	Moderate	4	30.8	22	57.9
	High	6	46.2	7	18.4

TEE—total energy expenditure, BMI—body mass index, PAL—physical activity level, n—number of participants, $\overline{\mathbf{X}}$—arithmetic mean, SD—standard deviation, BIM—beneficial immunomodulation, UBIM— unbeneficial immunomodulation, HUBIM—highly unbeneficial immunomodulation. Bold values denote statistical significance at the p < 0.05 level.

. At baseline, 13 participants (25.5%) were classified as BIM (diet supporting optimal immune function), 27 (52.9%) as UBIM (diet slightly weakening immune function), and 11 (21.6%) as HUBIM (diet significantly weakening immune function). For analysis, we combined UBIM and HUBIM into a single category (UBIM + HUBIM).

There were no significant differences between the BIM and UBIM + HUBIM groups in age, BMI, sleep duration, or physical activity level (PAL). Similarly, no significant differences were observed between groups in terms of diet type (vegetarian vs. traditional omnivore), marital status, education level, or smoking status.

Interestingly, a significant positive correlation was observed between body fat percentage and POLA index score (r = 0.40; p = 0.0041). The results are shown in Figure 1.

In the BIM group, at least 75% of participants met or exceeded the recommended intake levels for most individual dietary components, except vitamin D, iron, and calcium. In contrast, in the UBIM + HUBIM group, at least 75% of participants failed to meet the recommended intakes for linoleic acid, fiber, calcium, iron, folate, and vitamin D (

Table 2. Median (Q1–Q3) intake levels of selected nutrients in BIM vs. UBIM + HUBIM groups, expressed as percentage of recommended intake (RDA, AI, or EAR), with between-group comparisons.

Variable	BIM n = 13	UBIM + HUBIM n = 38	U Mann–Whitney Test
	Me (Q1–Q3)	Me (Q1–Q3)	p–Value
Water (% of adequate intake)	128.3 (121.5–181.3)	119.9 (98.2–144.6)	0.0631
Total protein	127.1 (103.6–153.8)	116.4 (100–135.3)	0.2520
Total fat	103.4 (92.1–118.4)	85 (69.1–97.7)	0.0038
Fats Total saturated	190.2 (164.8–206.6)	161 (135.9–209)	0.1476
Linoleic acid LA (C18:2)	114.1 (100.3–123.9)	76.2 (56.9–93.8)	0.0015
α-Linolenic acid ALA (C18:3)	134.3 (118.1–173.4)	91.7 (76.9–143.6)	0.0078
Assimilable carbohydrates	210.6 (192.9–231.2)	185 (155.2–203.2)	0.0196
Dietary fiber	126.8 (113–136.4)	79.6 (72.7–93.3)	<0.0001
Potassium	110.6 (98.3–134.9)	90.3 (75.9–103.2)	0.0023
Calcium	98.4 (64.9–106.3)	79.3 (61.2–90.8)	0.0601
Magnesium	132.6 (123.5–170.5)	103.5 (90.5–117)	<0.0001
Iron	102.5 (89.6–122)	70.4 (61.4–75.6)	<0.0001
Zinc	149.5 (122.9–172.3)	117.1 (99.2–126.1)	0.0019
Copper	212.6 (189.5–260.7)	150.8 (133.3–163.4)	<0.0001
Manganese	442.4 (364.3–475.2)	273.4 (231.1–354)	0.0001
Vitamin A	267.9 (203–307.5)	140.5 (118.3–189.5)	0.0003
Vitamin E (alpha-tocopherol equivalent)	182.6 (144.2–266.9)	127.3 (103.1–146.1)	0.0001
Thiamin	131.3 (115.5–195.2)	96.3 (76.7–110.8)	0.0003
Riboflavin	190.6 (165.6–242.7)	141.7 (118.4–164.1)	0.0021
Niacin	138.6 (113.2–192.6)	115.6 (88.1–155.5)	0.0800
Vitamin B6	244.6 (150.9–316.9)	125.8 (111.7–148.9)	0.0002
Folates	115 (105.3–123.9)	79 (65.8–87.5)	<0.0001
Vitamin B12	231.9 (163.7–506.9)	122.8 (93.7–172.6)	0.0065
Vitamin C	228.6 (189.4–334.1)	139.9 (92.1–233.6)	0.0146
Vitamin D	201.3 (35.7–311.6)	26.7 (16.1–92.6)	0.0246

n—number of participants, Me—median, Q1 and Q3—lower and upper quartile, BIM—beneficial immunomodulation, UBIM—unbeneficial immunomodulation, HUBIM—highly unbeneficial immunomodulation; bold p-values indicate statistical significance (p < 0.05).

).

Significant differences in the intake of certain food groups were observed between POLA index categories (

Table 3. Consumption of selected food products in BIM vs. UBIM + HUBIM groups. Continuous variables are presented as median (Q1–Q3) intake in grams per day; categorical variables represent the proportion of participants meeting predefined intake thresholds.

Variable		BIM n = 13		UBIM + HUBIM n = 38		U Mann–Whitney Test
		Me (Q1–Q3)		Me (Q1–Q3)		p-Value
Groats and rice [g/day]		14.5 (2.2–28.2)		9 (1.6–20.2)		0.6490
Seeds [g/day]		3.3 (1.7–11.4)		0.8 (0–3.1)		0.0110
Nuts [g/day]		18.3 (2.5–22.8)		11.5 (2.2–20.2)		0.4109
Seeds and nuts [g/day]		26.1 (15.4–36.8)		12.5 (3.8–23.1)		0.0743
Fruit [g/day]		242.9 (177.6–332.1)		199.7 (119–281.1)		0.1197
Vegetables [g/day]		439 (360.5–489.3)		308.1 (221.2–376.1)		0.0033
Total vegetables and fruit (in market products) [g/day]		663.4 (622.7–807)		474 (391.5–639.3)		0.0019
Legumes [g/day]		5.9 (0–45.5)		4 (0–10.7)		0.1980
Variable	Category	n	%	n	%	Chi2 p
Fruit and vegetables	<400 g/day	0	0.0	10	26.3	0.0001
	400–<600 g/day	1	7.7	15	39.5
	≥600 g/day	12	92.3	13	34.2
Nuts	10 g and more per day	9	69.2	21	55.3	0.2918
	Up to 10 g per day	4	30.8	17	44.7

n—number of participants, Me—median, Q1 and Q3—lower and upper quartile, BIM—beneficial immunomodulation, UBIM—unbeneficial immunomodulation, HUBIM—highly unbeneficial immunomodulation; bold p-values indicate statistical significance (p < 0.05).

). Participants in the BIM group consumed significantly more seeds and vegetables compared to those in the UBIM + HUBIM group. Regarding overall fruit and vegetable intake, 92.3% of BIM participants consumed ≥600 g/day, compared to only 42.1% in the UBIM + HUBIM group (p = 0.0017; Chi-square test). No significant differences were observed in legume or nut intake when analyzed separately, although combined intake of seeds and nuts tended to be higher in the BIM group.

3.2. POLA Index and Risk of COVID-19 or Influenza

The proportion of participants who reported COVID-19 or influenza during the follow-up period, compared to those who did not, stratified by POLA index category, is shown in Figure 2. In the BIM group, only 7.7% (1 of 13) reported infection after baseline, compared to 36.8% (14 of 38) in the UBIM + HUBIM group (p = 0.0446).

The association between POLA index category and the risk of self-reported COVID-19 or influenza is presented in

Table 4. Odds ratios for contracting COVID-19 or influenza during follow-up, comparing POLA index categories (UBIM + HUBIM vs. BIM) under varying levels of adjustment.

Model	OR (95% CI)		AR (95% CI)
POLA Index	BIM	UBIM + HUBIM	BIM	UBIM + HUBIM
Model 1 a	1 (ref.)	7.00 (1.18; 134.41)	7.7% (1.1%, 39.1%)	36.8% (23.2%, 53.0%)
Model 2 b	1 (ref.)	6.13 (0.84; 129.42)	8.3% (0.8%, 51.9%)	35.7% (14.5%, 64.5%)
Model 3 c	1 (ref.)	6.53 (1.02; 129.85)	12.3% (1.4%, 58.1%)	47.9% (26.1%, 70.6%)

^a—crude model; ^b—model adjusted for age, the type of diet, marital status, body fat percentage and smoking status; ^c—model adjusted for age and smoking; BIM—beneficial immunomodulation, UBIM—unbeneficial immunomodulation, HUBIM—highly unbeneficial immunomodulation, OR—odds ratio, AR—absolute risk.

. Participants in the UBIM + HUBIM group had markedly higher odds of illness compared to those in the BIM group. In the unadjusted model (Model 1), the OR for UBIM + HUBIM vs. BIM was 7.00 (95% CI: 1.18–134.41). This association remained elevated in the fully adjusted model controlling for potential confounders (Model 3: OR = 6.53; 95% CI: 1.02–129.85; p < 0.05 corresponding to a higher absolute risk of 47.9% (95% CI 26.1–70.6) versus 12.3% (95% CI 1.4–58.1) for UBIM + HUBIM vs. BIM). Although the association between immunomodulatory diet and infection risk lost statistical significance in some models, likely due to the limited sample size, the direction and magnitude of the effect remained consistent. In a more parsimonious model including only statistically significant explanatory variables, the BIM group had approximately 85% lower odds of infection compared with the UBIM + HUBIM group, although the confidence interval was wide.

3.3. Gut Microbiota

We next examined whether the POLA index category was associated with differences in gut microbiota composition, specifically focusing on bacteria with immunostimulatory potential (E. coli and Enterococcus spp., as cultured on selective media). No significant differences were observed between groups in the prevalence of E. coli and Enterococcus spp. above reference thresholds.

3.4. Mucosal Immunity Markers

Fecal markers of intestinal immunity (sIgA and HBD-2) and inflammation (calprotectin) were compared between POLA index categories. As shown in

Table 5. Fecal immunological markers in the BIM vs. UBIM + HUBIM groups. Values are presented as median (Q1–Q3).

Variable	BIM n = 13	UBIM + HUBIM n = 38	p-Value
	Me (Q1–Q3)	Me (Q1–Q3)
Calprotectin [µg/g]	9.07 (8.00–13.71)	12.29 (6.7–24.01)	0.2798
sIgA [µg/g]	1014 (349–1235)	1640 (1118–2990)	0.0125
β-defensin [ng/mL]	39.24 (13.88–88.21)	37.21 (19.53–128.06)	0.5890

n—number of participants, Me—median, Q1 and Q3—lower and upper quartile, BIM—beneficial immunomodulation, UBIM—unbeneficial immunomodulation, HUBIM—highly unbeneficial immunomodulation; bold p-values indicate statistical significance (p < 0.05).

, the median sIgA concentration was significantly lower in the BIM group than in the UBIM + HUBIM group (1014 µg/g vs. 1640 µg/g, p = 0.0125). Although the median levels of calprotectin and HBD-2 were also lower in the BIM group, these differences were not statistically significant.

We further evaluated the distribution of fecal sIgA levels relative to established laboratory reference ranges. As shown in Figure 3, POLA index category was strongly associated with sIgA status (p = 0.0412). In the BIM group, 92.3% of participants had an sIgA levels within the subnormal (<500 µg/g) or normal range (500–2040 µg/g), and only 7.7% had elevated sIgA levels (>2040 µg/g). In contrast, 39.5% of participants in the UBIM + HUBIM group had elevated sIgA levels.

These findings suggest that adherence to a diet with beneficial immunomodulatory properties (BIM) was associated with a more balanced mucosal immune profile, whereas lower-quality diets (UBIM + HUBIM) were more frequently associated with an elevated mucosal immune response.

3.5. Fecal Short-Chain Fatty Acids

Fecal SCFA levels were compared between groups, as presented in

Table 6. Fecal organic acid concentrations (mg/g dry stool) in the BIM vs. UBIM + HUBIM groups (median and interquartile range).

Variable	BIM n = 13	UBIM + HUBIM n = 38	p-Value
	Me (Q1–Q3)	Me (Q1–Q3)
Acetic acid	177.44 (121.5–220.0)	168.98 (142.13–209.98)	0.8121
Succinic acid	3.27 (1.26–3.81)	4.32 (2.49–19.24)	0.0422
Propionic acid	92.21 (59.47–113.0)	81.93 (63.28–100.83)	0.8968
Butyric acid	3.25 (2.59–4.16)	3.27 (2.15–5.91)	0.9223
Isovaleric acid	5.81 (3.55–21.69)	10.97 (4.72–17.96)	0.9785

n—number of participants, Me—median, Q1 and Q3—lower and upper quartile, BIM—beneficial immunomodulation, UBIM—unbeneficial immunomodulation, HUBIM—highly unbeneficial immunomodulation, bold values denote statistical significance at the p < 0.05 level.

. No significant differences were observed in the levels of most SCFAs, including acetate, propionate, butyrate, and isovalerate, between the BIM and UBIM + HUBIM groups. However, succinic acid levels were significantly lower in the BIM group compared to the UBIM + HUBIM group (median: 3.27 mg/g vs. 4.32 mg/g, p = 0.0422).

4. Discussion

4.1. POLA Index and Incidence of COVID-19 or Influenza

In this female pilot cohort, a lower POLA score (BIM category) was associated with a substantially lower self-reported incidence of COVID-19 or influenza over approximately 6 months of follow-up. Multivariable logistic regression models similarly indicated significantly higher odds of respiratory illness in the UBIM + HUBIM group compared to the BIM group. The direction and magnitude of this association are consistent with findings from our earlier report introducing the POLA index, which also demonstrated that higher POLA scores were associated with increased risk of viral respiratory infection [17]. These results reinforce the hypothesis that diets scoring favorably on the POLA index may confer protective effects against respiratory infections. Adequate dietary intake of anti-inflammatory and immunomodulatory nutrients plays a crucial role in supporting immune function and maintaining oxidative balance. This has become particularly relevant during the COVID-19 pandemic. For example, a dose–response meta-analysis by Hao et al. found that a higher DII was associated with increased incidence of SARS-CoV-2 [21]. Similarly, Aghajani et al. reported that greater adherence to Dietary Antioxidant Quality Score was significantly associated with reduced severity of COVID-19 infection [22]. However, it is important to interpret the current findings with caution. The wide CIs in our model reflect the limited number of infection events and potential model instability. Therefore, these results must be viewed as hypothesis-generating and require confirmation in larger, more adequately powered studies.

4.2. Gut Microbiota

We did not observe significant group differences in culture-based fecal counts of E. coli and Enterococcus spp. Interestingly, our earlier mixed-sex cohort study found an association between immunomodulatory dietary patterns and elevated E. coli counts in men, but not in a smaller subgroup of women [17]. In the present all-female cohort, no association was observed between diet quality (as captured by the POLA index) and E. coli colonization. It is important to note that culture-based methods provide only a limited snapshot of the gut microbial community. In future validation of the POLA index, gut microbiota analyses based on 16S rDNA sequencing or metagenome sequencing should be used to determine whether diets associated with favorable (low) POLA scores are related to beneficial changes in gut microbiota composition or function. This may then facilitate the identification of differences in the structure, diversity, and functional information of the gut microbiome between POLA groups. In addition, it will then be possible to calculate the Simpson index and Shannon index (reflecting microbial species diversity) and relate these results to POLA groups [23,24].

4.3. Markers of Intestinal Inflammation and Immunity

Secretory IgA is a key immune component of the gastrointestinal mucosal barrier. Its resistance to chemical degradation, along with its ability to cross-link and entrap microorganisms, enables it to function as a first line of defense in mucosal immunity. Secretory IgA prevents pathogen colonization and penetration of the intestinal barrier through a process known as immune exclusion [25,26]. Additionally, sIgA plays a regulatory role in shaping gut microbiota composition and maintaining intestinal homeostasis.

In our analysis, we observed significant differences in fecal sIgA concentrations across POLA index categories. Participants in the BIM group had sIgA levels within the normal range, while those in the UBIM + HUBIM group more frequently exhibited elevated sIgA levels. Fecal sIgA concentrations are known to fluctuate depending on diet quality and nutritional status [27]. Various dietary components, including vitamins, amino acids, fatty acids, polyphenols, oligo-/polysaccharides, and probiotics, contribute to maintaining optimal sIgA production and mucosal health. Recent studies have suggested that dietary interventions may not only modulate intestinal sIgA response but also influence interactions between sIgA and the gut microbiota [27,28]. Secretory IgA is primarily produced in dimeric form [29] by plasma cells located in the gut-associated lymphoid tissue, in response to antigens derived from food and microbes. Once secreted into the intestinal lumen, it interacts directly with the microbiota to support microbial balance and mucosal homeostasis [30].

In the UBIM + HUBIM diet group, elevated sIgA concentrations may reflect adaptive mucosal immune activation in response to dietary influences on the gut microbiota. In contrast, the presence of normal sIgA levels in the BIM group suggests a more balanced mucosal immune environment, likely supported by dietary patterns that promote gut homeostasis. Elevated sIgA can indicate either immune activation or effective mucosal defense. Therefore, we interpret ‘normal range’ sIgA as most consistent with intestinal homeostasis, recognizing that isolated levels of sIgA cannot, by themselves, distinguish protective from reactive elevations. This ambiguity reinforces the need to interpret sIgA in conjunction with the clinical context and other mucosal markers. In conclusion, maintaining microbiota diversity through appropriate nutrition and supporting optimal sIgA levels may enhance immune function and increase resistance to infectious diseases, including COVID-19.

4.4. Short-Chain Fatty Acids

Among the SCFAs and related organic acids measured, only succinic acid concentrations differed significantly between the POLA diet groups. Fecal succinate content was lower in the BIM group than in the UBIM + HUBIM group. No significant differences were observed for acetate, propionate, butyrate, or isovalerate. Emerging evidence indicates that succinic acid levels in fecal and blood samples may serve as biomarkers of gut dysbiosis—an imbalance in the intestinal microbial community—linked to altered microbial metabolic activity, particularly in succinate metabolism. For instance, patients with irritable bowel syndrome exhibit gut dysbiosis marked by an increased prevalence of succinate-producing bacteria, contributing to elevated succinate levels in both the intestinal lumen and bloodstream [31]. Under normal physiological conditions, fecal succinate concentrations typically range from 1 to 3 μM (or μmol/g); however, in patients with inflammatory bowel disease, concentrations as high as 7 to 25 mM have been reported [32]. The POLA index is strongly correlated with anti-inflammatory dietary effects, which may help explain the lower fecal succinate levels observed in the BIM group. This isolated difference could reflect subtle shifts in microbial metabolic pathways or host–microbiota interactions modulated by diet quality. However, given the small sample size and multiple comparisons, this finding should be interpreted with caution. Future studies should re-evaluate SCFA profiles in larger cohorts and consider composite markers (e.g., total SCFAs or butyrate:acetate ratios) while applying appropriate adjustments for multiple testing. Nonetheless, the observed association between a high-quality diet (BIM) and lower intestinal succinate is a potentially meaningful finding that warrants further investigation into its underlying mechanisms and clinical relevance.

4.5. Strengths and Limitations

Notable strengths of this study include the replication of an association between the POLA index and respiratory infection incidence in an independent cohort, as well as the comprehensive collection of data on diet, physical activity, gut microbiota, SCFAs, and mucosal immunity markers using harmonized methods. The POLA index itself represents a pragmatic and actionable tool for assessing diet quality, with a specific focus on immunomodulatory nutrients and food groups. However, several limitations should be considered when interpreting the findings. First, the sample size was small, and only 15 participants reported a COVID-19 or influenza event, resulting in wide CIs and reduced power to adjust for potential confounding variables. Second, the outcome was self-reported and composite in nature, lacking virologic confirmation and without distinction between COVID-19 and influenza. It is also possible that individuals classified into BIM and UBIM + HUBIM categories differed not only in dietary patterns but also in health-related behaviors or illness perception, which may have influenced the likelihood of self-reporting respiratory symptoms as COVID-19 or influenza. Third, the cohort was geographically limited to a single urban area and included only women aged 25–45 years, which restricts the generalizability of the results to other populations. Finally, the microbiota analysis relied on culture-based methods with a pragmatic threshold for E. coli counts; more sensitive molecular techniques might provide different additional insights into microbial community structure and function perspective.

Collectively, these limitations warrant caution in interpreting the findings and underscore the need for confirmatory research in larger, more diverse populations. The results should be considered exploratory and hypothesis-generating rather than definitive evidence of causality.

4.6. Implications and Future Directions

Despite its pilot nature, this study contributes to the growing body of evidence suggesting that adherence to a diet consistent with a favorable POLA index profile is associated with a reduced risk of self-reported respiratory viral infections and mucosal immune characteristics of intestinal homeostasis. From a methodological perspective, future studies should incorporate several key improvements: (a) differentiate COVID-19 from influenza using laboratory-confirmed diagnoses and clearly defined follow-up windows; (b) collect data on vaccination status and virus exposure, and, where feasible, perform time-to-event analyses; (c) evaluate POLA scores both categorically and as continuous variables (e.g., using spline models) to capture potential dose–response relationships; (d) integrate multi-omics microbiome profiling (e.g., 16S RNA or shotgun metagenomic sequencing) and targeted mucosal immune phenotyping to gain mechanistic insights; and (e) consider interventional designs (such as diet improvement trials) to directly test whether increasing an individual’s POLA score can causally reduce infection risk. These methodological refinements would enhance causal inference and help clarify the mechanisms linking diet quality, mucosal immunity, and susceptibility to respiratory infections.

5. Conclusions

In this prospective pilot cohort of 51 young adult women from Kraków, Poland, a favorable POLA dietary score (BIM category) was associated with a substantially lower self-reported incidence of COVID-19 or influenza over the follow-up period (7.7% vs. 36.8%). However, the small number of outcome events, reliance on self-report, and potential uncontrolled confounding mean that these findings are preliminary and should be considered as exploratory/hypothesis-generating. Future studies should validate the POLA index in larger and more diverse cohorts; use laboratory-confirmed, pathogen-specific outcomes with clearly defined risk periods; incorporate vaccination and exposure variables; apply analytical methods suited to few events (e.g., penalized regression); report absolute risk measures alongside ORs; and integrate advanced microbiome profiling with targeted mucosal immune assessments. If our observations are confirmed—especially through intervention trials—improving an individual’s POLA index could become a feasible component of strategies to prevent respiratory viral illnesses.