Abstract
Importance SARS-CoV-2 enters the nasopharynx to replicate; mechanical debridement with nasal irrigation soon after diagnosis could reduce morbidity and mortality.
Objective To determine whether initiating nasal irrigation after COVID-19 diagnosis reduces hospitalizations and death, and whether irrigant composition impacts severity.
Design Unblinded randomized clinical trial of two nasal irrigation protocols in outpatients PCR positive for SARS-CoV-2, nested in a prospective case:cohort using laboratory-confirmed cases in the CDC COVID-19 Case Surveillance dataset.
Setting Single-lab community testing facility associated with the emergency department (ED) in Augusta, GA.
Participants A consecutive sample of outpatients 55 years and older were contacted from daily COVID-19+ lab reports between September 24 and December 21 of 2020. Patients without supplemental oxygen use or cognitive barriers agreeing to same-day irrigation initiation were remotely consented. Among 826 screened, 321 were unable to be reached, 132 were ineligible, 294 refused participation, and 79 participants were enrolled.
Interventions Participants were randomly assigned adding 2.5 mL povidone-iodine 10% or 2.5 mL sodium bicarbonate to 240ml of isotonic nasal irrigation twice daily for 14 days.
Main Outcomes and Measures The primary outcome was hospitalization or death from COVID-19 within 28 days of enrollment by daily self-report confirmed with phone calls and hospital records, compared to the CDC Surveillance Dataset covering the same time. Secondary outcomes compared symptom resolution by irrigant additive.
Results Seventy-nine participants were enrolled (mean [SD] age, 64 [8] years; 36 [46%] women; 71% Non-Hispanic White). Analyzed by intention-to-treat, by day 28, COVID-19 symptoms resulted in 1/42 hospitalizations in those irrigating with alkalinization, 0/37 in the povidone-iodine group, (1.27%) and no deaths. Of nearly three million CDC cases, 9.14% were known to be hospitalized, with an additional 1.5% mortality in those without hospitalization data. The total risk of hospitalization or death (10.6%) was 8.4 times that of enrolled patients (SE=2.74; P=.006). 62 completed daily surveys (78%), averaging 1.8 irrigations/day. Eleven had irrigation complaints, and four discontinued. There were no significant differences by additive.
Conclusion SARS-CoV-2+ participants initiating nasal irrigation were over 8 times less likely to be hospitalized than the national rate.
Trial Registration ClinicalTrial.gov Identifier: NCT04559035
Author Approval All authors have filled out ICMJE and approved submission.
Conflict of Interest Statement Materials were provided by Neilmed Inc. and Rhinosystems Inc. The study was supported by funding from the Bernard and Anne Gray Donor Advised Fund Community Foundation for Greater Atlanta, Neilmed Inc., and Rhinosystems. No authors have conflict of interest.
Question After testing positive for COVID-19, will rapidly initiating nasal irrigation with alkaline or povidone-iodine isotonic solution reduce the risk of morbidity and mortality compared to a national dataset?
Findings In this randomized trial of 79 older adults nested in a case:control with the CDC COVID-19 National Dataset, 1.27% of participants initiating nasal irrigation were hospitalized or died, compared to 10.6%, a significant difference.
Meaning In older outpatients testing positive for SARS-CoV-2 who initiated nasal irrigation rapidly after diagnosis, risk of hospitalization or death was eight times lower than national rates reported by the CDC.
Background and Objectives
Mechanical binding of the SARS-CoV-2 spike protein-receptor to the ACE2 receptor is a target for pharmacologic and immunologic COVID-19 therapeutics. Genetic and clinical factors support an additional mechanical therapeutic option – nasal irrigation – may be uniquely effective. Sungnak et al. localized the necessary co-expression of ACE2 and TMPRSS2 primarily in the ciliated nasal epithelia.(1) The finding that viral “spike” protein mutations and N-Methylation of the nasal ACE2 receptor increased infectiousness further support a mechanical relationship between the viral particle and receptor.(2)
Ignaz Semmelweis pioneered handwashing to reduce infection in 1847. In emergency medicine and surgery, debriding infectious material with copious high-powered irrigation is standard practice. Nasal irrigation under pressure, or “nasal lavage”, has been demonstrated to reduce the duration and severity of both Coronaviridae and illnesses like flu with shorter incubation periods.(3) Irrigation should work best in clinical scenarios with long incubation and local non-hematogenous spread, where reducing viral load impacts severity. Otolaryngology reviews supported the concept and safety of nasal antimicrobials and irrigation.(4-6) If clinically effective, irrigation could be an inexpensive option rapidly available worldwide.
Given research supporting the virucidal activity of povidone-iodine against MERS and SARS-CoV-2(7-9) and the possible impact of alkalinization to reduce SARS-CoV-1 viral cell fusion and entry,(10) patients were randomized to add alkalinization or povidone-iodine to pressurized nasal lavage. We hypothesized rapid initiation of nasal irrigation after testing positive would reduce the severity of COVID-19. Our primary outcome was COVID-19 hospitalization or death, with secondary outcomes of symptom duration and severity.(11, 12)
Methods
Trial Design
The Rapid Initiation of Nasal Saline Irrigation (RINSI) study was an unblinded randomized clinical trial of alkalinized vs povidone-iodine nasal irrigation in outpatients aged 55 years and older recently PCR positive for SARS-CoV-2, nested in a prospective case:cohort using laboratory-confirmed cases in the CDC COVID-19 Case Surveillance dataset of patients 50 and older.(13) The trial protocol and statistical analysis plan appear in Supplement 1. The study was approved by the institutional review board at Augusta University in Augusta, Georgia. All participants provided remote informed consent in English.
Study Setting and Recruitment
This trial was conducted in Augusta, Georgia. Patients testing positive for COVID-19 by nasal swab or saliva PCR processed at a single lab at the Augusta University were recruited from September 24, 2020 to December 21, 2020. The 28-day follow-up was completed January 18, 2021. Signs at the testing site informed patients of the irrigation trial and eligibility criteria, and a recruitment flyer was given with testing. The daily laboratory-generated list of COVID-19 tests was screened for patient age, first positive test in the system, and location within 25 miles of Augusta University. Prospective participants were called consecutively between the hours of 9:00am through the early afternoon five to six days a week. On days where sufficient staffing for deliveries was a concern, the list was randomized for calling order. Participants interested in participation were assessed over the phone for inclusion criteria, and remote informed consent was completed per IRB policy.
Study materials were delivered to their residence by a member of the research team using COVID-19 precautions (masks, maintaining 6 ft. or more physical distance, door drop off) later that day. Materials consisted of a nasal irrigation device with 28+ accompanying saline pods/packets, two gallon jugs of distilled water, a physical copy of the consent form, an instructional sheet, and the randomly allocated additive (baking soda or povidone-iodine) with a 2.5ml scoop. The detailed instruction sheet contained directions to mix the irrigant materials, as well as links to a YouTube video demonstrating how to conduct irrigation with the relevant device. A previous study found nasal irrigation units under pressure (squeezing or pump) were superior to gravitational units.(14) In order to avoid bias toward any particular product, participants were assigned on alternate days to one of two pressure-based nasal irrigation systems (NAVAGE, Rhinosystems Inc.) or Neilmed Sinus Rinse (Neilmed Inc.). Patients were instructed to initiate irrigation on the same day of contact and enrollment. An investigator called the patient or their designated contact at day 2, 7, 14, and 28 to verify ED visits, hospitalization, or answer any questions.
Eligibility Criteria
Eligible patients had to be able to read the informed consent in English, agree to nasal lavage for 14 days with a 14-day follow-up, provide a back-up contact for clinical follow-up, and be available to receive materials and initiate irrigation that day. Exclusion criteria included current supplemental oxygen therapy, unwillingness to try nasal irrigation or current use of nasal irrigation, nasal surgery within the past year or chronic sinusitis, prior COVID-19 infection or positive test, symptoms longer than 7 days prior to testing, inability to complete surveys by computer or smartphone, and an allergy to iodine or shellfish. Hospital employees were initially excluded, given the unknown impact of greater risk due to exposure or lower risk due to T-cell immunity, but this exclusion criteria was ultimately removed.
Randomization
Patients were randomized to rinse with 240cc saline including 0.5 mL 10% povidone-iodine (0.1% final concentration) or 0.5 mL sodium bicarbonate twice daily for 14 days. Randomization was stratified by sex in 10 blocks of 10 random numbers by the first author using Random.org. With odd numbers signifying alkaline and even povidone-iodine, numbered opaque envelopes were prepared in separate sequences for male or female participants to be opened after consent, indicating the appropriate additive to be given to the patient.
Interventions and Masking
To avoid withholding effective treatment and to better assess adherence to nasal irrigation, all enrolled participants were given a pressurized nasal saline irrigation unit. Masking of study treatments (alkalinization or 0.1% povidone-iodine) was not undertaken. Study personnel reviewing hospital records to verify admission or death were masked to intervention.
Main Outcomes and Measures
The primary outcome was hospitalization for COVID-19 symptoms within 28 days of enrollment, by self-report and phone calls verified by the testing site hospital’s electronic medical records. Secondary outcomes in enrolled patients compared symptom resolution and home exposure, adherence to nasal irrigation, and any impact of antimicrobial or alkalization addition to the irrigant. Prompts to complete study materials were sent to participants via email from Qualtrics twice daily for the duration of the study period. To verify irrigation, patients uploaded pictures of used irrigation materials daily into the Qualtrics system.
In addition to demographic data, patients were asked preexisting medical history as found on the CDC person of interest form, including Chronic Lung disease (Emphysema, COPD), Asthma, Type 1 Diabetes Mellitus, Type 2 Diabetes Mellitus, Cardiovascular Disease, Hypertension, Chronic Renal Disease, weight and height to calculate obesity defined as BMI>30, Immunocompromised condition, and symptoms. Symptoms included the number of days since first subjectively sick, loss of smell, loss of taste, fatigue, presence or absence of fever >100.4°F, chills, muscle aches, runny nose, cough (new onset or worsening of chronic cough), shortness of breath, nausea or vomiting, headache, abdominal pain, and diarrhea.
Hospitalization and mortality data were compared to the National CDC Case Surveillance Public Use Dataset.(13) This dataset has 12 elements for all COVID-19 cases shared with CDC, including first positive specimen, first report to CDC, first day of illness, a summary “case earliest date”, laboratory-confirmed or suspected, symptom onset, demographic data, and a binary measure for pre-existing clinical conditions. Hospitalization, ICU, and mortality data have four options: yes, no, unknown (marked on form), or missing (nothing recorded). Following CDC research recommendations, we identified laboratory confirmed cases by “case earliest date” to match the testing dates in our sample. Between September 23, 2020 and December 21, 2020, 45% of laboratory-confirmed cases in the dataset included known hospitalization status. While a December Georgia Department of Public Health analysis indicated a minimum admission rate in age 50+ of 16.4% (K. Krohnert, email communication December 2020), this information was not published. For the same period, the COVID-19 Case Surveillance Public Use Data with Geography for Georgia-specific information had hospitalization data for only 5%, with a hospitalization rate of 50%, so the 12-item National CDC Case Surveillance dataset was selected as the most comprehensive and conservative comparison group.
Statistical Analysis
At the time we initiated the study, admission rates in older and African American communities were 40%,(15) and the population in our study environment was 50% Black. We assumed a baseline admission rate of 25%, with a large effect size of irrigation alone (Cohen’s d=7) and a smaller impact of the additive to the irrigant (Cohen’s d=3). To reduce hospitalizations by 60%, 100 patients per group would be required based on irrigant composition, and 79 based on irrigation versus control. After three months, the patient outcome data was evaluated and the admission and mortality rate in the patients enrolled was zero. Given the percentage of patients admitted in Georgia (16.4%), and the public health burden of COVID-19 the decision was made to stop enrollment and compare outcomes at the end of the observation period.
For patient characteristics between irrigant groups and devices used, t-tests were used unless non-normal distribution was noted, in which case the Mann Whitney U test was used. Baseline measures of duration of symptoms and ongoing reported symptoms by day were compared across the conditions. To test the effect of treatment group in symptom persistence, either t-tests for normally distributed data or the Mann-Whitney U test for non-normally distributed data was used.
The results are compared to laboratory-confirmed case data from the CDC.(13) Chi-square was used to evaluate differences in demographic proportions of sex, race, and age by 10-year tronche. We used an exact binomial test with Clopper-Pearson confidence intervals to compare observed hospital admission rates among participants compared with national rates of severe disease (admission or death) published by CDC. The exact binomial test is well-suited to assess the probability of observing the proportion of patients in this study. That proportion comparison is supplemented by an odds comparison across the study and CDC data (MedCalc Software Ltd. Odds ratio calculator. https://www.medcalc.org/calc/odds_ratio.php (Version 20.009)).
To avoid overestimating by reporting bias, we calculated the admission rate retaining cases where hospitalization status was missing or unreported in the denominator. Since only half of CDC cases included hospitalization status, our number should underestimate true national admissions. We included deaths only when hospitalization was unknown or missing as a rate over the total confirmed case denominator. Thus, our rate underestimates deaths. We did not include deaths in cases where hospitalizations are reported in our analysis to avoid counting outcomes with increased severity twice.
Results
During the study period, 826 unique patients aged 55 and older were eligible to be contacted within 24 hours of a positive PCR; 79 were able to be enrolled and receive irrigation materials on the day of contact.(Figure 1). After enrollment, 11 patients complained of discomfort or spotty epistaxis, with four discontinuing irrigation.
There was one COVID-19 related admission out of 79 patients assigned to nasal irrigation, 0/37 assigned to povidone-iodine and 1/42 patients in the alkalinization group (1.3%). One patient in the alkalinization group had a COVID-19 related ED visit but was not admitted. In addition to COVID-19 healthcare utilization, one patient reported an ED visit for a minor trauma, and one patient had a syncopal episode requiring admission for evaluation in the follow up period after resolution of COVID symptoms. These events were verified in the EHR database, and there were no additional ED visits or hospitalizations found in consented patients.
During the same enrollment period, in patients 50+ years the CDC reported 2937299 laboratory-confirmed cases with 268607 known hospitalizations, or 9.14% admission rate. There were 44,358 deaths where hospitalization was not reported, or 1.5%. Hospitalization or death occurred in for a total of 10.6% (OR:.125, 95%CI .017 to 0.9.) The CDC comparison rate of hospitalization and mortality was therefore 8.39 times higher than the hospital admission rate of patients enrolled in this nasal irrigation study (z[78]= -2.522; SE=2.74; P=.006). These reported data were not adjusted for underreporting of hospitalizations and disease cases.(16) The CDC group had a lower reported proportion of minority patients (Table 1), but 36.6% did not specify race. For the 785,285 CDC cases for whom both hospitalization and death were reported, 8.22% of patients expired. There were no deaths in our cohort.
Of the 79 enrolled, 57 patients completed the initial symptom and history questionnaire; patients reported a median of 3.3 days (IQR 2,5) of symptoms prior to enrollment. Twelve patients received their materials but didn’t record their first irrigation until the following day.
An online daily symptom and irrigation data collection survey was completed by 62 patients (median of 12 of 14 days, IQR 1,13.75). Of 631 daily online surveys, patients reported irrigating once per day (7.29%), twice daily (88.43%), or none (4.25%), averaging 1.79 irrigations per day. (Table 2) Patients were asked to take pictures of used irrigation materials to corroborate irrigation, but the number of used packets over time became difficult to assess for confirmation. Eleven patients had concerns with irrigation, four discontinued use.(Table 3)
Presenting symptoms present in over 50% of patients included fever, muscle aches, congestion, and headache. In other studies evaluating COVID symptoms, fatigue, headaches, anosmia and congestion persisted.(17) There were no statistical differences in symptomatic outcomes by irrigation unit used, (Figure 2) but symptom resolution of all or only one mild symptom among headache, fatigue, anosmia and congestion in 14 days was more likely in the povidone-iodine group (21/27) than the alkalinization group (17/35, p=0.0192).
Ten participants (12.7% by intention to treat) had household contacts who tested positive at least one day after enrollment, compared to 18.8% in a published meta-analysis.(11) There was no difference in risk of household spread by additive or irrigation unit. (Table 3)
Discussion
To the extent that our results generalize, pressurized nasal irrigation offers a safe and over the counter measure with potentially vital public health impact. Nationally, the reduction from 10.6% to 1.3% as of the month this writing (August 2021) would have corresponded in absolute terms to almost 780,000 fewer patients 55 and older requiring admission. Improved patient outcomes would be accompanied by corresponding reductions in pressure on ICU capacity as well as stress and risk to healthcare providers.
The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pathogen is a single stranded positive sense RNA virus, with a similar spike protein-receptor binding mechanism as SARS-CoV and MERS-CoV. Clinically COVID-19 differs notably from previous Coronaviridae: children are less impacted; obesity, diabetes, African American race and hypertension are independent risk factors; the relatively pathognomonic symptom of anosmia is present in up to 80% of patients,(18, 19) and the duration from infection to mild symptoms to severe symptoms in the alpha variant was prolonged.
The clinical differences in presentation support the current understanding that SARS-CoV-2 primarily enters and replicates in the nasopharynx, with olfactory neuroepithelium ACE2 expressed at 700 times the expression in lungs.(19, 20) Conditions increasing nasal ACE2 expression or sinus size (obesity, hypertension, pollution, older male sex) correlate with increased severity, further supporting targeting viral fusion in the nasopharynx.(21, 22) In contrast, populations lacking fully developed sinus area (children), with a high baseline practice of nasal irrigation (Laos, Vietnam), or higher mask compliance have decreased severity.(23, 24)
The size of the nasal cavity (and thus available ciliated epithelia) correlate with age and male sex.(23, 25) At a protein level, obesity and diabetes both increased expression of nasal ACE2 receptors, as did pollution and age.(26, 27) Together, the nasal cavity size, ACE2 expression and variolation explanation could account for lower pediatric severity and spread.(28) The degree of methylation of the ACE2 receptors (and thus stiffness and ease of viral attachment) is related to both race and epigenetic stress.(29, 30) Thus, increased virulence correlating with increased stability of the spike proteins in variants supports the mechanical hypothesis.
Given the local cell to cell rather than hematogenous spread and delay in activation of lung TMPRRS2,(31) the potential exists that mechanically debriding viral particles lodged in the ACE2 receptor, but not yet fused, can reduce severity. Furthermore, the variation in severity with methylation implies that not all particles are securely attached. The size variations in the entire nasal cavity, rather than just anterior nares, support the concept that full nasal cavity irrigation may be superior to nasal spray. Finally, the number of asymptomatic cases and the correlation of illness severity with viral load implied that even after PCR positivity, a window exists wherein lowering the viral load through irrigation could be clinically advantageous.
While nasal irrigation reduced symptoms of other Coronaviridae, flu,(3) and bacterial carriage in otolaryngology(32, 33), pathology from local spread and aspiration and the continued production of viral load locally suggest a potentially greater impact on COVID-19. Association of viral load with severity (24, 34, 35) suggests a different kind of cumulative pathology related to immune response, as well as the potential for reducing severity after the fact by debridement. Multiple studies have demonstrated immediate viral load reductions in vitro and in vivo with direct oral or nasal application of antivirals,(7-9) or the theoretical benefit of lavaging and gargling. (36-38) However, a small study of gargles and sprays of effective povidone-iodine did not show a significant reduction in viral load,(39) nor were clinical differences seen in a small randomized trial using twice daily nasal irrigation without an effective virucidal.(17)
Multiple commentaries have supported the concept and safety(37) of debridement.(4, 38, 40, 41) A host of investigators coming to the above-detailed conclusions initiated multiple prospective studies varying additives (Neem oil, ozone, surfactant, lactobacillis, virucidals), concentration, and chronicity for COVID prevention or treatment. Several of these ongoing studies in prevention have been limited by pandemic enrollment difficulties, or by statistical challenges from a lower-than-expected incidence in healthcare workers who may have T cell protection initially unanticipated during power analyses.
To our knowledge this is the largest prospective clinical trial using both twice daily large volume irrigation with a virucidal arm, and with documented adherence to irrigation. Moreover, the older and higher risk population in this study (with concurrent larger sinus area) may be most suitable to reducing morbidity and mortality.
Limitations
Our results support that irrigation, whether accompanied by alkalinization or an effective virucidal, reduces the likelihood of hospitalization. There are a number of limitations to our study design and execution.
The primary concern without a matched control group is the generalizability of our sample. The CDC database did not differ significantly by sex or age, but too many patients were missing race/ethnicity to meaningfully evaluate. The greatest risk of bias comes from preferentially reporting cases with hospitalization or death. Data suggests cases and hospitalizations are underreported rather than over,(16) however, and the CDC admission rate of 9.14% is lower than rates in other prospective studies with older populations. In a study of monoclonal antibodies delivered to outpatients testing positive, Chen et al found a 15% admission rate in patients 65+ or with BMI > 35.(42) Our sample came from a socioeconomically challenged catchment area, with average age 64 and BMI >30. In a similar health system to ours, Price-Haywood et al found a 39.7% admission rate; a Cochran database of minority patients’ admission rates in similar time periods and demographic location to our enrollment period consistently found admission rates as high as 60%.(15, 43)
While the goal was to initiate irrigation as quickly as possible after a positive test, healthcare infrastructure and testing turnaround time may limit the potential for fast intervention. The requirement to participate on 24-hour notice could have biased our sample toward healthier, technologically connected, higher socioeconomic status patients. The bias from feeling too well or too sick to participate could have been an issue, but approximately equal numbers declined participation due to “brain fog” as declined because they felt well and deemed any intervention unnecessary.
While irrigation could be an effective mechanical protection against variants in vaccinated people, adoption of a new hygiene intervention – or any intervention – is a barrier. Of the 537 patients contacted, 28 did not want to perform nasal irrigation. Of those who initiated irrigation, most continued twice daily use, but eleven had concerns about irrigating that were communicated to our staff. While only four discontinued irrigation, without the discussion and coaching adherence in the general population could be lower.
Finally, our data suggested that povidone-iodine might be superior to alkalinization. While most studies find low concentrations to be safe over months of application in younger patients,(6) studies using tenfold higher concentrations for gargling did find transient increases that resolved after the study.(39) For prolonged use, thyroid function testing in older individuals may be warranted.
Conclusion
As an intervention, pressurized nasal irrigation showed promise to reduce the severity of COVID-19 infection when initiated within 24 hours of a positive test. As large unvaccinated populations pressure evolution of variants, an effective mechanical outpatient intervention to reduce viral entry and hospitalizations can save lives and reduce the stress on hospital staff. Further research into the frequency and adjuvants of irrigation will be important not just for this pandemic, but for future viruses to come.
Data Availability
Data is available from the University of Augusta Department of Emergency Medicine Research Office, and online from the CDC COVID-19 Case Surveillance Public Use Data
https://data.cdc.gov/Case-Surveillance/COVID-19-Case-Surveillance-Public-Use-Data/vbim-akqf/data
Footnotes
The initial analysis evaluated hospitalization risk only for cases where the status was known. To reduce the risk of reporting bias, statistical analysis was revised to include all laboratory-confirmed cases in the denominator, and only deaths where hospitalization data was unreported. Figure 1 was revised to include only unique cases testing positive. Tables 1-2 were revised, and Table 3 was added to compare irrigation and device data. Authors with ICMJE forms were added.