This study was approved by the Research Ethics Committee of our institution under number 1166/2017. Written free and informed consent was obtained from all patients who agreed to participate.

All electronic medical records of patients treated in Rhinology Section of Sao Paulo Hospital (HU-HSP), Federal University of Sao Paulo (UNIFESP) - Escola Paulista de Medicina (EPM) between January 2013 and December 2019 were evaluated, and 378 treated patients with a confirmed diagnosis of CRS were selected. Of these, 257 patients had nasal topical corticosteroid treatment cycles with complete pre- and post-evaluation data and were included in this study, 205 patients with CRSwNP and 52 with CRSnNP.

• •

  Aged >18 years;

• •

  Confirmed diagnosis of CRS according to the American Academy of Rhinology criteria3;

• •

  Treatment cycles between 3 and 6 months of CSNI or CSNS;

• •

  Complete subjective and objective evaluation data, pre- and post-treatment cycle, between January 2013 and December 2019.

Use of oral, intramuscular or intravenous corticosteroids during the studied cycle.

A retrospective observational study (historical cohort) was carried out with two groups: CSNI and CSNS. The CSNI group had 108 patients who were submitted to at least one treatment cycle with CSNI (totaling 292 treatment cycles with complete data), which were compared to 149 patients who only underwent treatment cycles with CSNS (totaling 300 treatment cycles with complete data) in the CSNS group.

The treatment cycle was defined as a period of 3–6 months with topical nasal treatment with CS, with subjective and objective assessment at the beginning and at the end of this period recorded in electronic medical records. Patients in the CSNS group were treated with nasal CS in commercial spray formulations; while those in the CSNI group were treated with 1% compounded budesonide drops (each drop presented 500 µg of budesonide) or betamethasone cream (0.5 mg/g betamethasone) that were diluted in 250 mL of alkaline homemade saline solution (250 mL of water + 1 level teaspoon of table salt + 1 level teaspoon of sodium bicarbonate).

Epidemiological data such as gender and age were collected; associated diseases such as asthma, acetylsalicylic acid (ASA) intolerance, non-steroidal anti-inflammatory (NSAID) intolerance, dipyrone intolerance, NSAID-exacerbated respiratory disease (N-ERD), allergic fungal rhinosinusitis (AFRS); history such as current smoking, former smoking and previous endoscopic sinus surgery (ESS) for CRS; main complaint; and details on the topical therapy used (drug, dose, volume, time of use, adverse events).

The subjective assessment of symptom improvement during the cycle was followed by the investigation of adverse events, CRS exacerbations, need for surgery after the treatment, and systemic medications used.

The patients’ objective clinical evaluation was performed with the 22-item SinoNasal Outcome Test test (SNOT)-2210 and the Lund–Kennedy Endoscopic Score (LKES)11 applied pre- and post-treatment.

Sub-analyses were performed separating patients with CRSwNP and CRSnNP, previously submitted to ESS and those not submitted to ESS, in addition to sub-analyses comparing the alternative CSNI therapies (1% compounded budesonide drops and betamethasone cream).

Continuous variables were represented by mean and standard deviation. Categorical variables were represented by absolute frequencies (n) and percentages (%). Data distribution was analyzed using the Kolmogorov–Smirnov test. The Z-score was used to normalize the non-parametric distribution data. The comparison of continuous data was performed using Student’s t test. Pearson’s Chi-Square Test or Fisher’s Exact Test was used to compare the frequency between groups. The significance level was set at p ≤ 0.05. The SPSS version 24.0 software was used in the analyses.

The mean age of the total sample was 55.5 years (±13.2), with a predominance of women in patients with CRSnNP (69.2% vs. 49.3%; p = 0.01*). There was an uneven distribution between the groups regarding the main complaint, with a predominance of anterior rhinorrhea (30.1% vs. 14.2%; p = 0.005*) and facial pain (21.2% vs. 5.4%; p = 0.001*) in patients with CRSnNP, and a predominance of hyposmia (33.7% vs. 7.7%; p = 0.002*) in patients with CRSwNP. As for the comorbidities, there was an uneven distribution of the presence of asthma (42.4% vs. 9.6%; p < 0.0001*), ASA intolerance (14.6% vs. 1.9%; p = 0.01*), dipyrone intolerance (13.7% vs. 1.9%; p = 0.02*) and N-ERD (13.7% vs. 1.9%; p = 0.02*), with a predominance in patients with CRSwNP. The distribution data are shown in Table 1.

The initial patient assessment showed that the CRSnNP group had higher mean SNOT-22 scores than the CRSwNP group (39.5 vs. 29.6 points; p = 0.0003*). The questions with the greatest impact on the quality of life in the CRSnNP group were number 5 (Postnasal discharge; 2.53), number 1 (Need to “blow” nose; 2.32) and number 22 (Blockage/congestion of nose; 2.29), while in the CRSwNP group we identified number 21 (Sense of taste/smell; 2.54), number 5 (Postnasal discharge; 2.14) and number 1 (Need to “blow” the nose; 1.77). All questions showed an uneven distribution between the groups, except for questions 2, 3, 7 and 18.

The patients’ initial mean LKES was 4.7 with a standard deviation (SD) of 3.1. The CRSwNP group had higher scores, partly due to the presence of polyps that increase the LKES (5.0 vs. 3.0; p < 0.0001*). However, even removing the values ​​of polyps in the LKES, CRSwNP patients have significantly higher scores than CRSnNP (3.5 vs. 3.0; p = 0.04*). Patients with CRSwNP had a higher proportion of grade 2 (polypoid) edema (31.4% vs. 13.1%; p < 0.0001*); while those with CRSnNP showed a higher proportion of grade 2 (purulent) secretion (23.8% vs. 11.5%; p < 0.0001*). The pre-treatment impact of the disease (CRSwNP and CRSnNP) measured by the SNOT-22 and by the LKES can be analyzed in Table 2.

Considering now the treatment cycles, 592 cycles were counted, being 292 of CSNI and 300 of CSNS. In the 292 cycles of CSNI, the most often used corticosteroid was 1% compounded budesonide drops (51% of the cycles), followed by betamethasone cream (49%). The most often used daily dose of 1% compounded budesonide drops was 1000 µg (48.3%), followed by 500 µg (44.3%), diluted in 250 mL of saline solution. For the cream, 0.5 mg of betamethasone dipropionate (0.5 mg/g) was used, diluted in this same daily volume of saline solution. All patients used a syringe as an irrigation device. In the 300 cycles of CSNS, the most often used drug was budesonide (90% of the cycles), most of which (68.7%) at a dose of 400 µg per day. In addition, 89.7% of the CSNS cycles were accompanied by nasal irrigation with saline solution concomitantly.

There was a statistically significant difference in the epidemiological characteristics between patients who underwent cycles of CSNI and CSNS, with less female predominance (44.2% vs. 55.0%; p = 0.008*) and younger mean age (55.2 years vs. 57.7 years; p = 0.02*) in the CSNI group. Additionally, patients undergoing CSNI cycles had more comorbidities: dipyrone intolerance (15,4% vs. 9,0%; p = 0,02*) and N-ERD (14.7% vs. 9.7%; p = 0.06); and more co morbid antecedents: current smoking (12.0% vs. 3.7%; p = 0.0001*) and previous ESS (83.6% vs. 52.7%; p < 0.0001). There were no other initial differences, including the presence of nasal polyps (83.2% vs. 82.0%; p = 0.70).

The results of theCSNI and CSNS cycles in all patients are shown in Table 3. There was a significant improvement in LKES with both treatments (Irrigation: 4.8–4.4; p = 0.01*; spray: 4.6–4.3; p = 0.02*), but with more adverse events during CSNI cycles (2.4% vs. 0.3%; p = 0.04*). Ear fullness (2 patients), epistaxis (2), nasal irritation (1), epigastric pain (1) and nausea (1) were the adverse events that occurred with CSNI, while CSNS caused epistaxis in only one patient. CSNI showed to be more effective to prevent exacerbations than CSNI (6.5% vs. 10.7%; p = 0.07) but did not change the surgical indication rate after treatment (11.3% vs. 12.0%; p = 0.79).

When considering only CRSwNP, we obtained 489 treatment cycles, 243 of CSNI and 246 of CSNS. Patients with CRSwNP undergoing CSNI cycles showed a lower female predominance (41.2% vs. 51.2%; p = 0.03*), younger mean age (55.1 vs. 57.5; p = 0.04*), higher proportion of dipyrone intolerance (18.2% vs. 11.0%; p = 0.03*), of N-ERD (17.7% vs. 11.4%; p = 0.05*), higher proportion of current smoking (11.5% vs. 3.7%; p = 0.001*) and higher proportion of previous ESS (85.2% vs. 52.4%; p < 0.0001*) than those submitted to CSNS cycles.

The results of the CSNI and CSNS cycles only in patients with CRSwNP were shown in Table 4. There was a significant improvement in the LKES for both cycles (Irrigation: 5.1–4.8; p = 0.04* and spray: 4.9–4.6; p = 0.03*) with a tendency towards more adverse events (2.5% vs. 0.4%; p = 0.07) and fewer exacerbations (6.2% vs. 10 .6%; p = 0.08) in the CSNI group than in the CSNS.

There were only 103 cycles of treatments in patients with CRSnNP (49 cycles of CSNI and 54 of CSNS), with no difference regarding patient characteristics between the cycles, except for more patients with previous ESS in the CSNI group (75.5% vs. 53.7%; p = 0.02*). There were also no differences regarding the results between the cycles.

When considering only patients with previous ESS, 402 treatment cycles were analyzed, being 244 of CSNI and 158 of CSNS. Patients with previous ESS undergoing CSNI cycles showed a lower female predominance (40.6% vs. 51.3%; p = 0.04*), younger mean age (54.3 vs. 57.6; p = 0.008*) and more current smoking (13.9% vs. 5.7%; p = 0.009*), with no differences in the other parameters, including the presence of polyps (84.8% vs. 81.7%; p = 0.40*). As for the results of the cycles, patients with previous ESS undergoing CSNI cycles had a higher rate of subjective improvement (85.3% vs. 77.2%; p = 0.04*) and a lower rate of exacerbations (6.6 % vs. 13.3%; p = 0.02*) than those submitted to CSNS cycles, and only those submitted to CSNI cycles showed improvement in LKES (4.7–4.3; p = 0, 01*), as shown in Table 5.

In patients never submitted to ESS, 190 treatment cycles were performed (48 of CSNI and 142 of CSNS) and there were no differences regarding the characteristics of patients undergoing CSNI or CSNS cycles, including the presence of polyps (75.0% vs. 82.4%; p = 0.26), except for the lower rate of ex-smokers (6.3% vs. 19.0%; p = 0.04*). Only CSNI cycles patients showed significant improvement in the SNOT-22 (44.5–39.7; p = 0.04*), as shown in Table 6. On the other hand, only CSNS cycles patients showed significant improvement in LKES (5.4–4.7; p < 0.001*).

When considering the alternative modalities of CSNI (1% compounded budesonide drops and betamethasone cream), 292 cycles of CSNI were computed, divided into 149 cycles of 1% compounded budesonide drops and 143 cycles of betamethasone cream. There was no difference regarding patient characteristics between alternative CSNI therapies, except for a higher prevalence of asthma in the betamethasone cream group (26.9% vs. 38.5%; p = 0.03*) and current smoking in the group using 1% compounded budesonide drops (16.8% vs. 7.0%; p = 0.01*). Only 1% compounded budesonide drops showed a significant improvement in LKES (5.1–4.6; p = 0.02*), with fewer adverse events (0.0% vs. 4.9%; p = 0.006*), as shown in Table 7.

Considering only CRSwNP patients undergoing treatment with CSNI, which accounted for the largest number of CSNI cycles (83.2% – 243 cycles, with 119 cycles of 1% compounded budesonide drops vs. 124 cycles of betamethasone cream), this pattern was maintained, with improvement in LKES only in the group treated with 1% compounded budesonide drops (5.6–5.1; p = 0.02*) with fewer adverse events (0.0% vs. 4.8%; p = 0.03*). The same results were found when considering patients with previous ESS who underwent treatment with CSNI, also a large portion of CSNI cycles (83.6% – 244 cycles, being 119 cycles of 1% compounded budesonide drops and 125 cycles of betamethasone cream): improvement in LKES only in the 1% compounded budesonide group (5.0–4.6; p = 0.04*) with a lower rate of adverse events (0.0% vs. 5.6%; p = 0.02*).

A total of 72 cycles of CSNI with 1% compounded budesonide drops at a dose of 1000 µg and 66 cycles at a dose of 500 µg were analyzed. Patients with the higher dose (1000 µg) had more comorbidities: ASA intolerance (19.4% vs. 6.1%; p = 0.02*) and NSAID intolerance (13.9% vs. 4.6 %; p = 0.06). The higher dose (1000 µg) showed better results, with improvement in the SNOT-22 (30.7–27.0; p = 0.06) and improvement in LKES (5.5–4.8; p = 0.03*), as seen in Table 8.