← Back to Bird's Eye Reviews← Back to clinical trial results

Antiviral Nasal Sprays — Meta-Analysis

The top “across all compounds” plots compare every nasal spray on one outcome, standardized to Cohen's d (positive = favors treatment) so different scales and units pool together; each compound is a subgroup with its own subtotal. Below, one forest plot per ingredient × outcome × effect measure (≥2 trials), with a standardized plot leading each ingredient×outcome group. Pooled diamonds are random-effects (DerSimonian–Laird). (16 pooled.)

Across all compounds

Exploratory class-level pools — antivirals differ, so heterogeneity is high; read the prediction interval, per-compound subtotals, and the not-poolable appendix, not just the diamond.

Symptom duration — all compounds

Standardized (SMD)20 trials · 15 compounds · I² = 89.7%

Every compound pooled on one Cohen's-d scale (positive = favors treatment), combining continuous, hazard ratio, mean difference, odds ratio via Hedges' g, Chinn's OR→SMD, mean-difference→SMD (pooled-SD back-calculation) and an lnHR≈lnOR approximation. Pooling across different antivirals is exploratory — heterogeneity is high, so read the prediction interval, not just the diamond.

Compound / trialSMD (95% CI)Zanamivir (3)Monto 19990.14 (-0.00–0.27)Hayden 19970.04 (-0.26–0.33)Matsumoto 19990.70 (0.24–1.15)Subtotal0.23 (-0.06–0.51)Azelastine (2)Lehr 20250.24 (0.05–0.42)Meiser 20240.63 (0.34–0.91)Subtotal0.42 (0.04–0.80)Interferon alpha-2b (2)Kalyuzhin 20180.26 (-0.25–0.76)Turner 19860.81 (0.26–1.36)Subtotal0.52 (-0.02–1.07)Iota-carrageenan (2)Ludwig 2013-0.05 (-0.22–0.13)Hemilä 20210.38 (0.17–0.59)Subtotal0.16 (-0.26–0.58)Gel-based nasal spray (vicks first defence) (1)Little 2024-0.01 (-0.06–0.04)Hydrogen peroxide (1)Domênico 2021-0.01 (-0.28–0.27)Interferon beta serine (1)Sperber 19881.03 (0.16–1.90)Ivermectin (1)Aref 20211.82 (1.38–2.26)Nicotinic cholinergic agonist composition (1)Angulo 20210.24 (0.12–0.35)Pirodavir (1)Hayden 19950.07 (-0.31–0.45)Povidone-iodine (1)Polasek 2025-0.00 (-0.17–0.16)Povidone-iodine plus glycyrrhizic acid (1)Elsersy 20220.65 (0.36–0.93)Saline nasal irrigation (1)al. 20240.59 (0.00–1.17)Saline nasal spray (1)Little 2024-0.01 (-0.06–0.04)Saliravira (herbal compound) (1)Khorshiddoust 20221.02 (0.67–1.38)Overall (random effects)0.33 (0.21–0.45)-3.45-1.610.232.073.90← favors controlfavors treatment →

Heterogeneity: I² = 89.7%, τ² = 0.052 · 95% prediction interval -0.17 to 0.83

28 more studies reported this outcome but couldn't be standardized (shown, not pooled)
  • Douglas 1974 — Cp-20,961 (avridine) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Waldman 1978 — Cp-20,961 (avridine) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Farr 1984 — Interferon alpha (unspecified subtype) · Risk difference · no standardizable statistics (missing SD/CI/arm-N)
  • Hayden 1984 — Interferon alpha (unspecified subtype) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Frederick 1985 — Interferon alpha (unspecified subtype) · no standardizable statistics (missing SD/CI/arm-N)
  • Christian 1986 — Interferon alpha (unspecified subtype) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Monto 1986 — Interferon alpha-2b · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Hayden 1988 — Interferon alpha-2b · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Tannock 1988 — Interferon alpha-2a · no standardizable statistics (missing SD/CI/arm-N)
  • Sperber 1989 — Interferon beta serine · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Robert 1990 — Interferon (unspecified) · no standardizable statistics (missing SD/CI/arm-N)
  • Wiselka 1991 — Interferon alpha (unspecified subtype) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Sperber 1992 — Interferon beta serine · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Hayden 1996 — Zanamivir (gg167) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Adam 1998 — Hypertonic saline · Hazard ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Edward 2001 — Zinc sulfate · Hazard ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Susan 2002 — Zanamivir · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Eby 2006 — Zinc gluconate · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Hull 2007 — Hydroxypropyl methylcellulose (hpmc) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • James 2007 — Citrate/phosphate buffer · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Tamas 2012 — Iota-carrageenan · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Martin 2014 — Iota-carrageenan · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Eccles 2015 — Iota-carrageenan · no standardizable statistics (missing SD/CI/arm-N)
  • Ю.В. 2016 — Nazoferon · no standardizable statistics (missing SD/CI/arm-N)
  • Безрукова 2021 — Interferon alpha-2b · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Marcos 2022 — Chlorpheniramine maleate · no standardizable statistics (missing SD/CI/arm-N)
  • al. 2023 — A8g6 antibody cocktail · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Cristiana 2026 — Resveratrol/carboxymethyl-beta-glucan (linfovir) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)

Viral load — all compounds

Standardized (SMD)20 trials · 16 compounds · I² = 84.3%

Every compound pooled on one Cohen's-d scale (positive = favors treatment), combining continuous, hazard ratio, mean difference, odds ratio via Hedges' g, Chinn's OR→SMD, mean-difference→SMD (pooled-SD back-calculation) and an lnHR≈lnOR approximation. Pooling across different antivirals is exploratory — heterogeneity is high, so read the prediction interval, not just the diamond.

Compound / trialSMD (95% CI)Povidone-iodine (3)Kamal 20221.23 (0.40–2.06)Zarabanda 20220.37 (-0.47–1.21)Alsaleh 2024-0.09 (-1.28–1.10)Subtotal0.58 (-0.16–1.32)Azelastine (2)Meiser 20240.37 (0.12–0.62)Klussmann 20230.32 (-0.23–0.86)Subtotal0.36 (0.14–0.59)Nitric oxide (nons) (2)Tandon 20220.13 (-0.21–0.47)Winchester 20210.61 (0.16–1.06)Subtotal0.35 (-0.12–0.82)A8g6 antibody cocktail (1)al. 20230.01 (-0.32–0.34)Acidic hypochlorous solution (sentinox) (1)al. 2022-0.48 (-1.15–0.19)Astodrimer sodium (1)Winchester 20240.14 (-0.14–0.42)Baby shampoo (1)Esther 2022-0.70 (-1.28–-0.12)Hypertonic seawater solution with xylitol, panthenol, and lactic acid (tonimer lab panthexyl 800) (1)Cegolon 2022-1.10 (-1.87–-0.33)Iota-carrageenan (1)Eccles 20150.49 (-0.09–1.08)Ivermectin (1)Aref 20211.26 (0.86–1.66)Lentinan (1)Fan 20231.10 (0.24–1.96)Physiological seawater nasal irrigation (1)al. 2024-0.12 (-0.23–-0.02)Pirodavir (1)Hayden 19920.58 (-0.10–1.26)Ruprintrivir (1)Hayden 20030.74 (0.31–1.16)Zanamivir (1)Calfee 19990.29 (-0.58–1.16)Zanamivir (gg167) (1)Hayden 19961.30 (0.73–1.88)Overall (random effects)0.32 (0.09–0.56)-8.07-3.750.584.909.23← favors controlfavors treatment →

Heterogeneity: I² = 84.3%, τ² = 0.2075 · 95% prediction interval -0.68 to 1.33

56 more studies reported this outcome but couldn't be standardized (shown, not pooled)
  • Gatmaitan 1973 — Cp-20,961 (avridine) · no standardizable statistics (missing SD/CI/arm-N)
  • Douglas 1974 — Cp-20,961 (avridine) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Reed 1976 — Skf 40491, glr9-338, rp l9326, ici 73,602 · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Stanley 1976 — Cp-20,961 (avridine) · Risk difference · no standardizable statistics (missing SD/CI/arm-N)
  • Yasushi 1977 — 1-phenyl-3-(4-phenyl-2-thiazolyl) guanidine (cl 88,277) · Rate ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Aoki 1978 — Brl 5907 (polynucleotide interferon inducer of fungal origin) · no standardizable statistics (missing SD/CI/arm-N)
  • Stephen 1978 — Leukocyte/lymphoblastoid interferon · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Waldman 1978 — Cp-20,961 (avridine) · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Hayden 1982 — Enviroxime · Risk difference · no standardizable statistics (missing SD/CI/arm-N)
  • Levandowski 1982 — Enviroxime · no standardizable statistics (missing SD/CI/arm-N)
  • Hayden 1983 — Interferon alpha (unspecified subtype) · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Higgins 1983 — Interferon alpha (unspecified subtype) · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Phillpotts 1983 — Interferon alpha (unspecified subtype) · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Phillpotts 1983 — Enviroxime · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Samo 1983 — Leukocyte/lymphoblastoid interferon · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Hayden 1984 — Interferon alpha (unspecified subtype) · Risk difference · no standardizable statistics (missing SD/CI/arm-N)
  • Samo 1984 — Interferon alpha (unspecified subtype) · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Douglas 1985 — Interferon alpha (unspecified subtype) · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Higgins 1986 — Interferon beta serine · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Turner 1986 — Interferon alpha-2b · no standardizable statistics (missing SD/CI/arm-N)
  • Al-Nakib 1987 — 4',6-dichloroflavan · no standardizable statistics (missing SD/CI/arm-N)
  • HJORDIS 1987 — Interferon alpha (unspecified subtype) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Frederick 1988 — Rhinovirus receptor murine monoclonal antibody (rrma) · no standardizable statistics (missing SD/CI/arm-N)
  • Hayden 1988 — Interferon alpha-2b · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Sperber 1988 — Interferon beta serine · no standardizable statistics (missing SD/CI/arm-N)
  • Dearden 1989 — R61837 · no standardizable statistics (missing SD/CI/arm-N)
  • Sperber 1992 — Interferon beta serine · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Frederick 1995 — Pirodavir · Risk difference · no standardizable statistics (missing SD/CI/arm-N)
  • Hayden 1996 — 4-guanidino-neu5ac2en (gg167) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Frederick 1997 — Zanamivir · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Jack 2002 — Interferon alpha-2b · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Susan 2002 — Zanamivir · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • James 2007 — Citrate/phosphate buffer · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Ron 2010 — Iota-carrageenan · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Tamas 2012 — Iota-carrageenan · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Martin 2013 — Iota-carrageenan · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Martin 2014 — Iota-carrageenan · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Zhang 2014 — Ribavirin · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Ю.В. 2016 — Nazoferon · no standardizable statistics (missing SD/CI/arm-N)
  • Kalyuzhin 2018 — Interferon alpha-2b · no standardizable statistics (missing SD/CI/arm-N)
  • Charles 2021 — Baby shampoo · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Yilmaz 2021 — Hypertonic alkaline nasal irrigation · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Hazem 2022 — Povidone-iodine plus glycyrrhizic acid · Risk ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Julie 2022 — Povidone-iodine · no standardizable statistics (missing SD/CI/arm-N)
  • Marcos 2022 — Chlorpheniramine maleate · no standardizable statistics (missing SD/CI/arm-N)
  • Reza 2022 — Saliravira (herbal compound) · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Rujipas 2022 — Povidone-iodine · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Shankar 2022 — Povidone-iodine · no standardizable statistics (missing SD/CI/arm-N)
  • Batioglu-Karaaltin 2023 — Povidone-iodine / hypertonic alkaline / saline nasal lavage · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Caroline 2023 — Iota-carrageenan · no standardizable statistics (missing SD/CI/arm-N)
  • Rui 2023 — Igm-like ace2 fusion protein (hh-120) · Hazard ratio · no standardizable statistics (missing SD/CI/arm-N)
  • Song 2023 — Igm-like ace2 fusion protein (hh-120) · no standardizable statistics (missing SD/CI/arm-N)
  • Zhang 2023 — A8g6 antibody cocktail · no standardizable statistics (missing SD/CI/arm-N)
  • Francesca 2024 — Inna-051 · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Peter 2024 — Povidone-iodine · Mean difference · no standardizable statistics (missing SD/CI/arm-N)
  • Thomas 2025 — Povidone-iodine · Mean difference · no standardizable statistics (missing SD/CI/arm-N)

By compound

iota-carrageenan

Quality of lifeStandardized (SMD)Std. mean difference · 2 trials · I² = 43.4%

Standardized effect size (Cohen's d): combines continuous & odds ratio onto one scale via Hedges' g and Chinn's OR→SMD conversion (an approximation). Positive = favors treatment. 2 trials converted from binary.

TrialStd. mean difference (95% CI)Eccles 20150.06 (-0.22–0.34)Eccles 20100.56 (-0.13–1.25)Pooled (random)0.21 (-0.24–0.66)-0.360.070.500.931.37no effect (difference = 0)

iota-carrageenan

Symptom durationStandardized (SMD)Std. mean difference · 2 trials · I² = 89.2%

Standardized effect size (Cohen's d): combines hazard ratio & mean difference onto one scale via Hedges' g and Chinn's OR→SMD conversion (an approximation). Positive = favors treatment. 2 trials converted from binary.

TrialStd. mean difference (95% CI)Ludwig 2013-0.05 (-0.22–0.13)Hemilä 20210.38 (0.17–0.59)Pooled (random)0.16 (-0.26–0.58)-0.32-0.080.170.410.65no effect (difference = 0)

iota-carrageenan

Symptom severityStandardized (SMD)Std. mean difference · 3 trials · I² = 50.2%

Standardized effect size (Cohen's d): combines continuous & mean difference onto one scale via Hedges' g and Chinn's OR→SMD conversion (an approximation). Positive = favors treatment. 3 trials converted from binary.

TrialStd. mean difference (95% CI)Eccles 20150.24 (-0.04–0.52)Unger-Manhart 2024-0.10 (-0.53–0.34)Eccles 20100.74 (0.02–1.46)Pooled (random)0.22 (-0.15–0.58)-0.69-0.110.461.041.62no effect (difference = 0)

iota-carrageenan

Symptom severityMean difference · 2 trials · I² = 64.3%
TrialMean difference (95% CI)Unger-Manhart 20240.02 (-0.19–0.24)Eccles 2015-0.61 (-1.32–0.10)Pooled (random)-0.20 (-0.79–0.39)-1.46-0.96-0.470.030.52no effect (difference = 0)

zanamivir

Influenza symptom durationStandardized (SMD)Std. mean difference · 3 trials · I² = 67.7%

Standardized effect size (Cohen's d): combines continuous & mean difference onto one scale via Hedges' g and Chinn's OR→SMD conversion (an approximation). Positive = favors treatment. 3 trials converted from binary.

TrialStd. mean difference (95% CI)Monto 19990.14 (-0.00–0.27)Hayden 19970.04 (-0.26–0.33)Matsumoto 19990.70 (0.24–1.15)Pooled (random)0.23 (-0.06–0.51)-0.380.040.450.861.27no effect (difference = 0)

zanamivir

Influenza symptom durationMean difference · 3 trials · I² = 0%
TrialMean difference (95% CI)Matsumoto 1999-1.00 (-1.62–-0.38)Hayden 1997-1.00 (-1.80–-0.10)Monto 1999-1.00 (-2.00–0.00)Pooled (random)-1.00 (-1.45–-0.55)-2.16-1.58-1.00-0.420.16no effect (difference = 0)

povidone-iodine

Symptom severityStandardized (SMD)Std. mean difference · 2 trials · I² = 0%

Standardized effect size (Cohen's d): combines continuous & mean difference onto one scale via Hedges' g and Chinn's OR→SMD conversion (an approximation). Positive = favors treatment. 2 trials converted from binary.

TrialStd. mean difference (95% CI)Polasek 20250.16 (-0.10–0.42)Alsaleh 20240.04 (-1.14–1.23)Pooled (random)0.16 (-0.10–0.41)-1.33-0.650.040.731.42no effect (difference = 0)

povidone-iodine

Viral loadStandardized (SMD)Std. mean difference · 3 trials · I² = 46.9%

Standardized effect size (Cohen's d): combines continuous & odds ratio onto one scale via Hedges' g and Chinn's OR→SMD conversion (an approximation). Positive = favors treatment. 3 trials converted from binary.

TrialStd. mean difference (95% CI)Kamal 20221.23 (0.40–2.06)Zarabanda 20220.37 (-0.47–1.21)Alsaleh 2024-0.09 (-1.28–1.10)Pooled (random)0.58 (-0.16–1.32)-1.54-0.580.391.362.32no effect (difference = 0)

azelastine

Symptom durationStandardized (SMD)Std. mean difference · 2 trials · I² = 80.4%

Standardized effect size (Cohen's d): combines continuous & hazard ratio onto one scale via Hedges' g and Chinn's OR→SMD conversion (an approximation). Positive = favors treatment. 2 trials converted from binary.

TrialStd. mean difference (95% CI)Lehr 20250.24 (0.05–0.42)Meiser 20240.63 (0.34–0.91)Pooled (random)0.42 (0.04–0.80)-0.070.190.460.720.98no effect (difference = 0)

azelastine

Viral loadStandardized (SMD)Std. mean difference · 2 trials · I² = 0%

Standardized effect size (Cohen's d): combines continuous onto one scale via Hedges' g and Chinn's OR→SMD conversion (an approximation). Positive = favors treatment. 2 trials converted from binary.

TrialStd. mean difference (95% CI)Meiser 20240.37 (0.12–0.62)Klussmann 20230.32 (-0.23–0.86)Pooled (random)0.36 (0.14–0.59)-0.310.000.320.630.94no effect (difference = 0)

nitric oxide (nons)

Viral loadStandardized (SMD)Std. mean difference · 2 trials · I² = 64.2%

Standardized effect size (Cohen's d): combines continuous & mean difference onto one scale via Hedges' g and Chinn's OR→SMD conversion (an approximation). Positive = favors treatment. 2 trials converted from binary.

TrialStd. mean difference (95% CI)Tandon 20220.13 (-0.21–0.47)Winchester 20210.61 (0.16–1.06)Pooled (random)0.35 (-0.12–0.82)-0.310.060.420.791.16no effect (difference = 0)

nitric oxide (nons)

Viral loadMean difference · 2 trials · I² = 51%
TrialMean difference (95% CI)Tandon 2022-0.52 (-0.92–-0.12)Winchester 2021-1.21 (-2.07–-0.35)Pooled (random)-0.76 (-1.40–-0.11)-2.24-1.64-1.03-0.430.17no effect (difference = 0)

interferon alpha-2b

Symptom durationStandardized (SMD)Std. mean difference · 2 trials · I² = 52.4%

Standardized effect size (Cohen's d): combines continuous onto one scale via Hedges' g and Chinn's OR→SMD conversion (an approximation). Positive = favors treatment. 2 trials converted from binary.

TrialStd. mean difference (95% CI)Kalyuzhin 20180.26 (-0.25–0.76)Turner 19860.81 (0.26–1.36)Pooled (random)0.52 (-0.02–1.07)-0.380.090.551.021.49no effect (difference = 0)

interferon beta serine

Symptom severityStandardized (SMD)Std. mean difference · 2 trials · I² = 42.9%

Standardized effect size (Cohen's d): combines continuous onto one scale via Hedges' g and Chinn's OR→SMD conversion (an approximation). Positive = favors treatment. 2 trials converted from binary.

TrialStd. mean difference (95% CI)Sperber 19890.21 (-0.02–0.44)Sperber 19880.80 (-0.05–1.65)Pooled (random)0.36 (-0.15–0.87)-0.290.230.751.281.80no effect (difference = 0)