Aim of the study

We conducted a pre-post clinical study to evaluate with a comprehensive approach whether a standard cycle of nasal inhalations with thermal water was able to influence the level of inflammatory biomarkers, ameliorate short-term symptoms, rhinomanometry measurements, and clinical objective evaluation in adult and elderly patients with chronic rhinitis, pharyngitis or sinusitis. The confounding effect of cigarette smoking, age, occupational exposures and diet was taken into account.

Secondary aims of the study were to evaluate the compliance to the treatment, the occurrence of side-effects, and the persistence of beneficial effects of thermal water nasal inhalations on patients’ symptoms 1 year later. In addition, the potential interaction of the patients’ clinical, demographic, socio-economic and lifestyle profile with the effect of the treatment was investigated.

Patients recruitment

The study population included adult and elderly consecutive patients with chronic rhinitis, pharyngitis or sinusitis (International Classification of Disease - tenth revision (ICD10): J30 vasomotor and allergic rhinitis, J31 chronic rhinitis, nasopharyngitis and pharyngitis, J32 chronic sinusitis) admitted at Terme di Genova spa (Genoa, Italy) for nasal inhalations with thermal water, male and female.

Approval has been obtained by the competent Ethics Committees of IRCCS San Raffaele Pisana, Rome, and of the Liguria Region.

“Nasal inhalations with thermal water” or “inhalation treatment” mean a cycle of combined warm vapor inhalation and aerosol by nasal adapter for 12 days in 2 weeks with “Cappelletta” source water. The mineral water of this source, which is located in Acquasanta (Genoa, Italy), has an appreciable hydrogen sulfide degree and is a salty earthy alkaline water, due to a prevalence of sodium, calcium and chloride ions, with a pH higher than 11 on average. This water has been used for the treatment of respiratory diseases since ancient times.

Exclusion criteria were systemic or topic treatment within the last 2 week with mucolytics, corticosteroids, anticongestants, antihistamine or NSAIDs, severe chronic diseases, acute disease or surgery in the last month, previous thermal treatment less than 12 months before.

Conditions which determined exclusion from the study were interruption of the treatment before completion, receiving less than 10 sessions, treatment lengthening for more than 4 days, and a new diagnosis of acute or chronic severe disease, hospital admission or surgery during the study period.

All eligible patients admitted to the spa received detailed information about the study by the physician in charge of the treatment, and thereafter were asked to sign an informed consent. All patients who accepted to participate to the study and signed the consent were requested to donate biological samples and were administered a questionnaire.

As is shown in Fig. 1, 233 eligible patients meeting the inclusion/exclusion criteria were asked to participate to the project and 183 (78.5%) entered in the study (T0), while 50 refused, mainly due to time constraints.

Among the patients enrolled, 18 (10%) discontinued the treatment and were considered as dropouts. In sensitivity analysis, they did not show any statistically significant difference from the others neither in terms of diseases of interest nor smoking habits (Pearson χ² test), nor as regards total symptom score (TSS) and clinical evaluation score (CES) at enrolment. On the other hand, they were significantly younger on average (49.4 ± 15.3 vs. 61.8 ± 15.3, p = 0.001).

Fifteen patients, out of the 165 who completed the treatment and underwent the final evaluations (T1) did not fully respect the protocol, e.g. extension of the end-of-treatment date was longer than 4 days (generally due to personal reasons or to the difficulty to reach the spa, which was located in the countryside). Finally this small group was kept in the analysis, because in sensitivity analysis treatment efficacy was not statistically different when comparing these subjects to the whole study group in terms of TSS, CES and CXCR4 expression, although they were significantly younger (Kruskal-Wallis test).

The first 79 patients enrolled were called on the phone 12 months after the end of the treatment (T2), to administer a questionnaire on their upper respiratory symptoms. Only 74 replied while 5 of them could not be reached (6.3% lost to follow up).

Data and sample collection

A trained physician visited each patient at admission (T0), filling a custom structured questionnaire to collect socio-demographic information and data on job (including exposure to inhalable noxious substances), lifestyle (including smoking habit, exposure to second-hand smoke and diet), clinical history (general and ORL, both remote and specifically referring to the last year), symptom score and clinical evaluation score. The symptom questionnaire was a simplified version of that used in the Naiade survey [30], and was based on 10 items, including cough, nasal obstruction, and nasal discharge with 4 degrees of intensity (0 to 3). TSS were calculated for each subject by summing all the scores assigned to each symptom in the scale. CES were calculated assigning a score 0 to patients without pathologic signs and a score 1 to patients with any pathologic features referring to nasal/pharyngeal mucosa or secretion (e.g.: normal, red, pale or hypertrophic nasal mucosa in left and right nostril; maximum possible CES: 6).

During clinical evaluation, a sample of nasal epithelial cells was collected with a rhinoprobe from the two nostrils for inflammatory biomarkers measurement. An active anterior rhinomanometry was performed in order to measure nasal resistance in a subgroup of patients.

After 2 weeks, at the end of the inhalation therapy (T1), nasal specimen and data on symptoms, clinical evaluation, and rhinomanometry were collected again. The patient was asked about side-effects.

One year after the end of the study (T2), the first 50% of patients enrolled were searched by phone and interviewed using the structured questionnaire for symptom scale, and they were requested to report on any major changes in domicile, job, smoking habit and diet.

Specimen tubes and questionnaires were labeled with unique numerical codes, representing the identity codes for each subject, ensuring blindness of the study and respect of privacy.

A customized database was developed to store patients and specimen information. Access to personal data was restricted.

The samples were stored at − 30 °C no longer than 2 weeks, then transferred at − 80 °C and periodically shipped in dry ice to the laboratory in Rome for analysis.

Additionally, we have included in the study 10 healthy subjects, to ensure a proper comparison of inflammatory biomarker levels with the patients affected by upper respiratory tract diseases. These subjects were enrolled in the Department of Sense Organs of the Sapienza University of Rome and did not suffer from allergic rhinitis, chronic rhinitis, sinusitis, nasopharyngitis and pharyngitis nor were under anti-inflammatory therapies. They were 7 men and 3 women with a mean age of 50 ± 6 years, non-smokers and without any major inflammatory event in the 3 months preceding sampling of nasal epithelial cells.

Real-Time Quantitative Polymerase Chain Reaction (RT-PCR)

Samples of nasal epithelial cells were lysed to extract total RNA with the Ambion Paris system (Life technologies). After isolation, approximately 1 μg of RNA was reverse transcribed using the High-Capacity cDNA Archive Kit (Applied Biosystems, Milan, Italy) following manufacturer’s instructions. Aliquots of cDNA were subjected to real time PCR in 50 μl of 1 x Universal PCR Master Mix, 0.5 μM TaqMan probe and 5 ng of cDNA. Primers and probes for HIF-1α, NF-kB, P2X7R, RAGE, NOS2, COX2, PTX3, CXCR4, VEGF, etc. (see Additional file 1: Table S1) were designed using Assays on-Demand facility (Applied Biosystems). Each sample was loaded in duplicate, and negative and positive controls were included. Amplification of 18S rRNA was used as internal reference gene. PCR amplifications were performed using an ABI PRISM 7900 sequence detector (Applied Biosystems). Amplification data were analyzed using the Sequence Detector version 1.7 software (Applied Biosystems). Statistical analysis of real-time PCR results were done using mean normalized cycle threshold (ΔC_t) values and the pooled standard deviation of the mean ΔC_t.

Western blot assay

Samples were lysed to extract total proteins with the Ambion Paris system (Life technologies). Lysates were clarified by centrifugation (10 min at 4 °C) and the supernatant collected. Protein concentration was determined by the Bradford assay (Bio-Rad). Equivalent amounts of protein were electrophoresed on SDS-polyacrylamide gels. The gels were then electroblotted onto PVDF membranes. After blocking with 5% milk, membranes were incubated with the primary antibody overnight. Finally, the relevant protein was visualized by staining with the appropriate secondary horseradish peroxidase-labeled antibody for 1 h followed by enhanced chemiluminescence.

The following primary antibodies were used: mouse anti-CXCR4, rabbit anti P2X7R, goat anti-AGER, rabbit anti-SOCS1 (Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-PTX3 (Alexis Biochemical, San Diego, U.S.A.) and mouse anti-ACTB (Sigma-Aldrich, St Louis, MO). The following secondary antibodies were used: mouse anti-rabbit HRP, goat anti-mouse HRP (Amersham Biosciences, Piscataway, NJ), donkey anti-goat HRP (Santa Cruz).

Statistical analysis

The primary endpoint of the study was the assessment of the safety and efficacy of thermal water nasal inhalations administered to the patients diagnosed with chronic rhinitis, pharyngitis or sinusitis. Measures of effect included changes of symptoms and clinical conditions, rhinomanometry measurements, and inflammatory biomarkers levels from T0 to T1 (primary outcome) and T2 (only for symptoms; secondary outcome). The intent-to-treat (ITT) set included all subjects who were enrolled in the trial. The per protocol (PP) population was the subset of the ITT set that included subjects who completed the study with no major protocol violations.

Descriptive analyses were performed to report on the distribution of clinical and epidemiological variables among the patients, while the variation of the frequency of study endpoints within study intervals was evaluated with the paired samples t-tests (a “repeated measures” t-test). Whenever the distribution of parameters did not allow the use of parametric statistics, non parametric tests such as Mann-Whitney U Test or Wilcoxon signed rank sum test were used instead. In some cases log-transformation was applied. Statistical significance was set at p < 0.05 and appropriately adjusted by Bonferroni correction for multiple comparisons when required. Statistical analyses have been conducted using STATA/SE version 12·1 statistical software and MedCalc for Windows, version 15.0 (MedCalc Software, Ostend, Belgium).

Patients profile

Nine percent had a low education level, while 15% had a university degree. Half of the patients were retired and 12% housewives, while more than one third held an employment.

As regards the pathologies of interest, half of the patients suffered from sinusitis (ICD10 code: J32), 29% from (naso) pharyngitis (J31.1-J31.2), a quarter had chronic, vasomotor or allergic rhinitis (J30, J31.0), with 10 subjects declaring two pathologies at once. The proportions of each disease differed according to age: among the most aged (≥65 years), sinusitis ranked first, followed by rhinitis and pharingitis (42, 35 and 28% respectively), while the youngest were affected mostly by sinusitis (60%), then pharingitis (30%) and rhinitis (14%). The percentages of subjects with two diseases at once were similar (5–6%).

Seventy-one percent of the subjects declared no allergy and 16% a respiratory allergic condition.

Current substantial contact with potentially noxious inhalable substances during work or leisure was reported by a minority of patients (n = 5), and for 13% this exposure was considered sporadic or low. Fifty percent of patients never smoked cigarettes, while 14% were current smokers. A low percentage of the subjects declared to be exposed to second-hand smoke at work or at home (7%).

Half of the patients had normal weight, while 38% were overweight and 10% obese. The proportion of subjects eating two portions or less of fruit and vegetables per day was limited to 14%, with 44% of the sample in the intermediate category and 42% declaring to consume at least 6 portions per day. One subject out of four consumed vitamin or antioxidant dietary supplements.

Total symptom score and clinical evaluation score

Symptoms reported by the patients in T0, T1 and T2 are illustrated in Fig. 2. The percentage of patients who declared to suffer with nasal obstruction, nasal discharge, scratchy sore throat, etc. was slightly diminished after treatment among the subjects with low intensity symptoms (panel A) and strongly decreased among those who had symptoms with an intermediate to strong intensity (panel B). The average TSS was 5.7 ± 3.3 at admission, and was reduced by more than 50% after the inhalation treatment (2.7 ± 2.3, statistically significant difference), as is shown in Table 2A. No major changes were found after stratification for age, for pathology or for smoking habit (not shown). Symptoms remained at substantially low levels at T2, that is at the phone follow up 1 year after treatment (mean TSS 2.3 ± 2.6).

A) Total symptom scores in T0 - T1 - T2, and clinical examination scores in T0 – T1
	T0 Mean ± SD (range) N = 183		T1 Mean ± SD (range) N = 165	T2 Mean ± SD (range) N = 74	P - valuea (T0 vs. T1)		P - valuea (T0 vs. T2)
Total symptom score (TSS)	5.7 ± 3.3 (0–18)		2.7 ± 2.3 (0–12)	2.3 ± 2.6 (0–12)	< 0.00001		< 0.00001
Clinical examination score (CES)	1.4 ± 1.2 (0–5)		0.6 ± 0.7 (0–4)	n.a.	< 0.0001		n.a.
B) Rhinomanometry measurements in T0 - T1
	Inhalation resistance (Pa/ml s)				Exhalation resistance (Pa/ml s)
	N	T0 Mean ± SD	T1 Mean ± SD	P - valueb	N	T0 Mean ± SD	T1 Mean ± SD	P - valueb
All patients	105	0.40 ± 0.35	0.51 ± 0.61	NS	103	0.40 ± 0.35	0.53 ± 0.60	=0.01
Pathologyc
J30-J31.0 Rhinitis	28	0.40 ± 0.38	0.63 ± 1.00	NS	28	0.40 ± 0.32	0.67 ± 0.96	NS
J31.1-J31.2 Pharyngitis	24	0.43 ± 0.43	0.44 ± 0.37	NS	24	0.46 ± 0.41	0.45 ± 0.26	NS
J32 Sinusitis	57	0.38 ± 0.28	0.49 ± 0.44	NS	55	0.36 ± 0.31	0.49 ± 0.43	=0.009
Cigarette smoke
Never smokers	47	0.39 ± 0.32	0.44 ± 0.38	NS	49	0.39 ± 0.34	0.46 ± 0.38	NS
Former smokers	46	0.39 ± 0.34	0.53 ± 0.11	NS	41	0.37 ± 0.33	0.60 ± 0.74	NS
Currentsmokers	12	0.52 ± 0.45	0.69 ± 0.84	NS	13	0.48 ± 0.40	0.62 ± 0.81	NS

n.a. not assessed, SD standard deviation, NS not statistically significant (p > 0.05), SD standard deviation

^aWilcoxon signed rank sum test

^bWilcoxon signed rank sum test T0 vs. T1

^cFigures do not add up because 10 patients declared two pathologies simultaneously

At admission, the patients underwent an objective clinical examination: about one quarter of the subjects had deviated septum, 4 out of 10 had pathologic signs in their nasal or pharyngeal mucosa (41 and 38% respectively), 9% had nasal hypersecretion and one patient had pharyngeal hypersecretion. The percentage of patients with any pathological signs in the nasal or pharyngeal mucosa decreased to less than 50% in T1 as compared to T0, while similar but less pronounced patterns were shown for nasal or pharyngeal secretion, which were basically observed only in a minority of the subjects. More in detail, after treatment 17% of the patients had pathologic signs in their nasal mucosa and 14% in their pharyngeal mucosa, while nasal or pharyngeal hypersecretion was found in 4 and 11% of the patients, respectively. Globally, mean CES was 1.4 ± 1.2 at admission and 0.6 ± 0.7 after treatment (statistically significant difference). Results were unchanged when stratifying for age or pathology, while never smokers had lower values with respect to current or former smokers (not shown).

Rhinomanometry

Table 2B shows the main results of active anterior rhinomanometry. Total nasal resistance was globally higher after inhalation treatment (statistically significant in the case of exhalation resistance). The differences between T0 and T1 inhalation and exhalation resistances were correlated (Spearman, rho = 0.74, p < 0.0001). After stratification by pathology, a statistically significant variation was shown only for exhalation resistance among patients with sinusitis, while the values in T0 and T1 were almost identical among subjects affected with pharyngitis.

No statistically significant difference between measurements in T0 and in T1 remained after stratification of the sample by smoking habit, however never smokers had the lowest values, and current smokers the highest. In addition, increases in T1 tended to be more pronounced among men than women, among elderly patients compared to other adults, and for strong consumers of vegetables and fruits with respect to low or medium consumers (data not shown).

Gene and protein expression

A panel of 48 genes implicated in the innate immunity system and inflammation were analyzed in nasal epithelial cells with High Throughput real-time PCR with a dedicated mini-wells card. This analysis was preliminarily conducted in 40 patients randomly selected among those who reached T1, to identify the candidate gene (s) to be considered for further analyses, among those whose expression was most altered after the inhalation treatment.

After stratifying for pathology, we found that CXCR4 expression was consistently changed in patients suffering from all three diseases of interest, while FLT1 demonstrated a statistically significant post-treatment reduction only in those affected by sinusitis. Other genes showed weaker changes, although in some cases the change reached statistical significance, namely SOD2 average expression was more than double after treatment among patients with pharyngitis (p = 0.007), TLR4 was one third lower in subjects affected with rhinitis (p = 0.02), and NFKB1 was 20% higher in the subgroup with sinusitis (p = 0.04).

When stratifying by smoking habit (data not shown), the increase of CXCR4 mean expression in T1 was confirmed in all subgroups, while FLT1 was significantly altered only among non-smokers. The small subgroup of current smokers (n = 4) showed a somehow peculiar profile, with highly significant changes in the expression of ICAM1 and NFKB1 (p < 0.0001 and p = 0.002 respectively) and some other minor variations.

Finally, we decided to focus on CXCR4 in further analyses, considering the notable difference in absolute figures between mean expression in T0 and in T1, the high statistical significance and the consistency in subgroups, while the other genes were discarded, also taking account of the multiple comparison issue.

CXCR4 protein was visualized by Western blot of the cell lysate with the appropriate antibodies in the same 40 patients plus 10 reference subjects (Fig. 3 and Table 4). The mean value of the patients’ CXCR4 expression at baseline was similar to that of the controls and was 2.5-fold increased after inhalation treatment, confirming the results obtained with High Throughput real-time PCR, although no statistical correlation was demonstrated between the two endpoints. The same results were found in the three pathologies of interest or among never smokers and former smokers even after stratifications. On the other hand, the mean value of CXCR4 expression at baseline resulted to be higher in the small subgroup of current smokers than in the non-smokers (p = 0.05), and the increase after treatment was smaller and non significant.

	CXCR4 protein expression				CXCR4 gene expression
	N	T0 Mean ± SD	T1 Mean ± SD	P - valuea	N	T0 Mean ± SD	T1 Mean ± SD	P - valuea
Controls	10	0.97 ± 0.51		NSb	10	0.66 ± 0.44	–	= 0.025b
All patients	40	1.0 ± 0.39	2.6 ± 0.66	< 0.0001d	143	1.2 ± 0.68	3.3 ± 1.2	< 0.0001d
Pathologyc
J30-J31.0 Rhinitis	10	1.2 ± 0.37	2.4 ± 0.66	< 0.0001d	34	1.1 ± 0.6	3.5 ± 1.2	< 0.0001d
J31.1-J31.2 Pharyngitis	7	1.2 ± 0.45	2.7 ± 0.6	= 0.0001d	37	1.2 ± 0.59	3.1 ± 1.3	= 0.0001d
J32 Sinusitis	24	0.96 ± 0.37	2.6 ± 0.69	< 0.0001d	77	1.1 ± 0.77	3.3 ± 1.1	< 0.0001d
Cigarette smoke
Never smokers	23	0.99 ± 0.39	2.6 ± 0.66	< 0.0001d	68	1.2 ± 0.66	3.5 ± 1.2	< 0.0001d
Former smokers	13	1.0 ± 0.36	2.6 ± 0.50	< 0.0001d	56	1.1 ± 0.57	3.3 ± 1.2	< 0.0001d
Current smokers	4	1.4 ± 0.38	2.4 ± 1.2	NS	19	1.1 ± 0.98	2.7 ± 0.93	< 0.0001d

NS not statistically significant (p > 0.05), SD standard deviation

^aFisher’s exact Test for difference T0 vs. T1

^bDifference controls vs. T0

^cFigures do not add up because some patients declared two pathologies simultaneously

^dDifference statistically significant after Bonferroni adjustment for multiple comparisons (cut-off is ≈ 0.005)

Results were substantially unchanged when comparing aged subjects (≥ 65 years) or the eldest quartile (≥ 71 years) to the others. No major changes were shown after stratifying for gender, allergy, exposure to second-hand smoke, exposure to noxious inhalable substances or fruit and vegetables consumption.

Finally, expression levels of the gene CXCR4 were measured through RT-PCR in all the available samples (143 patients and 10 controls; Table 4). The results obtained with the first 40 patients were substantially confirmed in this larger group, with a 2.75-fold increase of the mean values in T1 respect to T0, a highly statistically significant increase. In addition, CXCR4 gene expression at baseline was significantly augmented in patients compared to controls.

The considerable increase after treatment was substantially maintained after stratifying for the three groups of pathologies or according to smoking habit. The increase was smaller in the subgroup of current smokers than in non-smokers (CXCR4 gene expression level in T1: 2.7 ± 0.93 vs. 3.4 ± 1.2, p = 0.02).

Results were substantially unchanged after stratifying for age, gender, allergy, exposure tosecond-hand smoke and exposure to noxious inhalable substances.

Associated events

Among the 163 patients who completed the inhalation treatment, 7 (3.8%) suffered from the so-called “thermal crisis” (a temporary exacerbation of their symptoms lasting 1–2 days, which is considered a systemic reaction to the treatment and has no pathologic meaning [31]), and 13 (7.1%) reported other negative events (e.g. headache, fatigue).