Use of Platelet-rich Plasma for COVID-19 Related Olfactory Loss, A Randomized Controlled Trial

Abstract

Introduction:

This study evaluated the use of platelet-rich plasma (PRP), an autologous blood product with supraphysiologic concentrations of growth factors, in the treatment of prolonged COVID-19 related smell loss.

Methods:

This multi-institutional, randomized controlled trial recruited COVID-19 patients with objectively measured smell loss (University of Pennsylvania’s Smell Identification Test, UPSIT≤33) between 6-12 months. Subjects were randomized to 3 intranasal injections of either PRP or sterile saline into their olfactory clefts. Primary outcome measure was change in Sniffin’ sticks score (TDI) from baseline. Secondary endpoint measures included responder rate (achievement of a clinically significant improvement, ≥5.5 point TDI), change in individual TDI olfaction scores, and change in subjective olfaction via a visual analogue scale.

Results:

35 subjects were recruited and 26 completed the study. PRP treatment resulted in a 3.67 point (95% CI: 0.05-7.29, p=0.047) greater improvement in olfaction compared to the placebo group at 3-months and a higher response rate (57.1% versus 8.3%, odds ratio 12.5, 95% exact bootstrap CI 2.2-116.7). There was a greater improvement in smell discrimination following PRP treatment compared to placebo, but no difference in smell identification or threshold. There was no difference in subjective scores between PRP and placebo. No adverse effects were reported.

Conclusion:

Olfactory function following COVID-19 can improve spontaneously after 6 months and can improve to a greater extent with PRP injection. This data builds on the promise of PRP to be a safe potential treatment option for patients with COVID-19 smell loss, and larger-powered studies will help further assess efficacy.

Introduction

Persistent post-viral olfactory dysfunction (OD) caused by the COVID-19 pandemic is a prominent global health concern that continues to rise.¹ While many achieve spontaneous recovery from their OD², persistent smell and associated taste loss are common symptoms of COVID long hauler syndrome with significant impacts on quality of life.^3–5

Potential therapies for post-viral and COVID-19 related OD remain limited with low efficacy and a paucity of evidence-based support. While a Cochrane review last updated December 2020 found no definitive treatments for persistent COVID-19 OD⁶, there are multiple ongoing clinical trials and a few recently published studies. The strongest evidence for the treatment of COVID-19 OD recommends olfactory training.^7–11 Other proposed therapies based on pre-pandemic evidence involve the use of topical intranasal medications^12–14 and oral anti-inflammatory / neuro-protective agents.^15–18 However, the efficacy remains moderate at best amongst all current proposed therapeutics.

This study evaluated the use of platelet-rich plasma (PRP), an autologous blood product with supraphysiologic concentrations of growth factors, in the treatment of prolonged COVID-19 related smell loss. PRP is widely used in other clinical fields and has demonstrated promise in peripheral nerve regeneration through stimulation of vascular and axonal regeneration via growth factors and by regulation of inflammatory response in the microenvironment.¹⁹ This study builds from prior pilot studies published by our group and others that demonstrated the safety of PRP in its use for OD.^20–22 In a murine model of anosmia, topical intranasal PRP resulted in improved olfactory function and restoration of an intact olfactory epithelium.²³ Prior single-arm clinical trials utilizing PRP for OD demonstrated no adverse outcomes and a potential improvement in olfactory function.^20,21 Notably, Steffens and colleagues recently demonstrated the potential efficacy of a single intranasal injection of PRP for the treatment of COVID-19 related OD compared to olfactory training. Though that study had a limited follow-up period and lacked randomization or blinded placebo arm, its results build upon our group’s pilot data that suggests PRP may play a role in the treatment of post-viral OD. The aim of this randomized controlled clinical trial was to evaluate the efficacy and safety of intranasal PRP in a cohort of subjects with COVID-19 related persistent OD despite mainstay treatments including olfactory training.

Methods

This study was a randomized, single-blinded, placebo-controlled trial comparing the use of PRP with sterile saline intranasal injection in participants with persistent COVID-19 induced OD. The study was approved by the Stanford University (IRB#55353) and the University of California San Diego (IRB#210296) Institutional Review Board Committees and registered on Clinicaltrials.gov (NCT04406584).

Participant Selection

Participants were recruited from patients seen in the Rhinology clinics at Stanford and UCSD between June 2021 and May 2022 who had PCR-confirmed diagnosis of SARS-CoV-2 between April 2020 and October 2021, and objective OD duration of > 6 months but < 12 months as depicted by the study flow diagram in Figure 1 according to CONSORT guidelines. Six months duration was used as a cutoff to ensure that the majority of patients known to spontaneously improve after COVID-19 induced smell loss would not confound improvement from the intervention.²⁴ One year duration was used as a cutoff as we know the duration of loss often predicts recovery prognosis²⁵ and potentially how well any intervention may benefit our smell loss patients and we did not want to miss a significant finding based on extended duration. Inclusion criteria comprised of adult patients >18 years of age with confirmed olfactory dysfunction with a quantitative score of ≤33 points on the University of Pennsylvania’s Smell Identification Test (UPSIT) prior to study randomization. Participants must have previously trialed both olfactory training⁸ and topical budesonide nasal irrigations¹² for at least 3 months and have a normal endoscopic exam of the nasal cavity and olfactory cleft. Exclusion criteria comprised of a history of inflammatory sinonasal disease or evidence of rhinitis or sinusitis on endoscopy, prior sinonasal or anterior skull base surgery, self-reported OD prior to SARS-CoV2 infection, neurodegenerative disease, history of bleeding disorders or the use of blood thinner medications.

Figure 1:

CONSORT flow diagram of the study recruitment and analysis

Outcome Measures

The outcome instrument was Sniffin’ Sticks²⁶ a validated olfactory psychophysical test to determine odor threshold, discrimination, and identification (TDI) with each component score ranging from 0-16 for a total possible score of 48. Primary outcome measure was change in TDI score from baseline. Secondary endpoint measures included responder rate at 3 months, where a responder was defined as a clinically significant improvement on Sniffin’ Sticks TDI score (≥5.5pts). Additional secondary endpoints were the change in individual TDI component scores from baseline, and subjective olfaction via a 0-10 point visual analog scale (0 = no smell, 10 = perfect smell).

Study design

Subjects underwent 1:1 randomization to either PRP or placebo (sterile saline) treatment via a random number generator. All recruited participants were initially screened for olfactory dysfunction using UPSIT score ≤33 and then underwent baseline olfactory psychophysical testing using Sniffin’ Sticks. Repeat Sniffin’ Sticks testing was performed at the 4-week (1-month) and 3-month follow-up visits. Subjective smell function was queried at each time point.

Prior to treatment, participants were topically anesthetized with pledget application of 4% lidocaine and 0.1% phenylephrine. Subjects received 1ml of either PRP or sterile saline injected submucosally into bilateral olfactory clefts under endoscopic visualization. Treatments were given 2 weeks apart at 3 different time points (week 0, week 2, and week 4). All participants were blinded to the treatment received, underwent phlebotomy, and wore a blindfold during injections.

PRP Preparation and Injection

PRP isolation and injections were performed as depicted (Supplemental Figure 1) and previously described in our pilot study.²⁰ Emcyte GS30-PURE II PRP kits were utilized and PRP isolation was performed per GS30-PURE II Protocol A (Emcyte, Ft Myers, FL). Of note, PRP kits were donated by the Emcyte corporation, but the study design, completion, and data analysis were conducted solely by the authors. In brief, 25ml of whole blood was obtained through a peripheral blood draw and added to a pre-filled syringe with 5mL of sodium citrate anticoagulant. The sample was centrifuged for one minute at 4200rpm, the platelet plasma suspension was aspirated and re-centrifuged for 5 minutes at 4200rpm. The subsequent supernatant containing platelet poor plasma was discarded leaving 2.5ml of PRP that was resuspended and drawn up into 2 separate 1ml sterile syringes and injected submucosally at two sites within the olfactory cleft along the superior septum, posterior to the head of the middle turbinate. Participants in the placebo study arm received 1ml sterile saline injections bilaterally in the same locations.

To confirm proper isolation of PRP, whole blood and PRP samples from select participants (n=9) were processed for complete blood count analysis. Compared to their respective whole blood, PRP samples resulted in an average 5.9-fold increase in platelet concentration (Supplemental Figure 1) with low granulocyte and red blood cell counts.

Statistical analysis

We performed a power analysis based on data from our pilot study in which hyposmic participants with a mean baseline olfaction score (Sniffin’ Sticks) of 22.4 points and a standard deviation of 4.6 points improved 5.85 points following PRP therapy.²⁰ Thus we determined that a sample of 20 participants (10 control, 10 experimental) would provide this trial with 80% power to detect a similar effect size at 26% improvement at 3 months, at a two-sided alpha level of 0.05.

Shapiro-Wilk test was used to confirm the Gaussian distribution of TDI scores, TDI component scores and subjective olfaction scores. To compare patients’ baseline demographic and clinical characteristics between two study arms, Fisher’s exact test was used for discreet variables and t-test for continuous variables. Linear mixed regression models were used to determine the effect of PRP and placebo interventions on olfaction scores over the 1-month and 3-month trial period, because such models avoid listwise deletion of an entire study participant and thus yield unbiased estimates when missing data occurred at a particular timepoint. First degree of autocorrelation covariance structure was chosen for all the models as it yields the best Bayesian Information Criterion (BIC) model fitting score. An interaction term between study arm and study month was included in the model to compare the differences in change of olfactory scores. The model also controlled for baseline olfactory score.

At 1 month and 3 month timepoints, we calculated the responder rate, or the percentage that achieved a minimally clinically important difference (MCID) in TDI score, previously determined as an improvement of ≥5.5 points. Due to our small sample size, we opted to use the median unbiased estimate of the probabilities of MCID to estimate the odds ratios at month 1 and month 3 and calculated 95% confidence intervals based on “exact” bootstrap distribution.²⁷

SAS software version 9.4 (SAS Institute) was used to perform statistical analyses. A p-value <0.05 (2-sided) was considered significant.

Results

This multi-institutional single-blinded RCT assessed 35 subjects for eligibility; 29 of which completed the trial through month-1(n=17 intervention, n=12 placebo), and 26 of which completed the month-3 trial (n=14 intervention, n=12 placebo, Figure 1). Five subjects did not meet eligibility criteria (4 tested normosmic by Sniffin’ sticks despite testing hyposmic on UPSIT screening and 1 had a history of smell loss due to prior trauma). Of the 30 subjects who were randomized, 1 in the PRP arm failed to complete intervention (disqualified with new diagnosis of a bleeding disorder / severe thrombocytopenia). Three additional subjects in the PRP arm completed the 1-month follow-up, but were excluded from the 3-month analysis due to loss to follow-up, recurrent COVID-19 infection, and nasal surgery within the follow-up period. Baseline characteristics and clinical demographics for the participants were similar between the two study arms, as reported in Table 1. The mean age of recruited subjects was 44.1 years (SD 14.0 years) and 50% were female. There were no differences in the average duration of OD (placebo 8.6 months vs. PRP 8.9 months, p=0.73). Baseline olfactory scores between placebo and PRP arms were similar as measured by UPSIT (25.2 vs. 22.4, p=0.28) and Sniffin’ Sticks (26.0 vs. 24.3, p=0.41). As part of the inclusion criteria, all subjects had OD for at least 6 months following their COVID-19 infection and had previously trialed olfactory training and high volume topical nasal steroid rinses without resolution of their OD.

Table 1:

Participant Demographics

	Mean (SD)
Mean (SD)	Placebo n=12	PRP n= 18	p-value
Age, yrs	43.4 (16.3)	44.6 (12.7)	0.832
Gender (n,% male)	6 (50.0%)	9 (50.0%)	1.000
Duration of Olfactory loss, mo	8.6 (2.4)	8.9 (2.2)	0.725
Parosmia (%)	5 (41.7%)	13 (72.2%)	0.101
Subjective Smell, 0-10	3.8 (2.0)	3.9 (1.4)	0.876
Baseline UPSIT Score	25.2 (6.9)	22.4 (6.7)	0.282
Baseline Sniffin’ Sticks (TDI) Score	26.0 (4.4)	24.3 (6.4)	0.413
Baseline T Score	5.0 (2.7)	5.0 (3.9)	0.975
Baseline D Score	10.7 (1.7)	9.5 (2.6)	0.186
Baseline I Score	10.4 (2.8)	9.8 (2.6)	0.533
Race (n,%)			0.446
Hispanic	2 (16.7%)	6 (33.3%)
White, Non-Hispanic	9 (75.0%)	9 (50.0%)
Black, Non-Hispanic	0 (0.0%)	1 (5.6%)
Two or more races	1 (8.3%)	1 (5.6%)
Asian or Pacific Islander	0 (0.0%)	1 (5.6%)
American Indian or Alaskan Native	0 (0.0%)	0 (0.0%)
PMH (n,%)
Diabetes	0 (0.0%)	1 (5.6%)	0.424
Hypertension	1 (8.3%)	1 (5.6%)	0.775
Asthma	2 (16.7%)	2 (11.1%)	0.674
Allergies	1 (8.3%)	1 (5.6%)	0.775
Long-hauler symptoms (n,%)
Shortness of breath	0 (0.0%)	0 (0.0%)	-
Fatigue	1 (8.3%)	2 (11.1%)	0.812
Headache	0 (0.0%)	0 (0.0%)	-
Palpitations	0 (0.0%)	1 (5.6%)	0.424
Brain fog	0 (0.0%)	2 (11.1%)	0.247

SD-standard deviation, UPSIT- University of Pennsylvania Smell Identification Test, TDI- threshold, discrimination, identification

Using a linear mixed model that adjusted for baseline score, estimated mean improvement in objective (TDI) and subjective (VAS) olfactory function are summarized for both placebo and PRP arms at 1-month and 3-months post-intervention in comparison with baseline (Table 2). The PRP arm had a statistically significant improvement above baseline Δ4.31 TDI points, 95% CI: 1.69-6.93 at 1-month post-intervention (p=0.002) and Δ6.25 points, 95% CI: 3.85-8.65 at 3-months (p<0.0001). The placebo arm had no statistically significant improvement above baseline (Δ1.17, −1.99-4.32 and Δ2.58, −0.13-5.29) at 1- and 3-months respectively. Examining individual components of olfaction (Table 2): threshold (T), discrimination (D), and identification (I) all improved post-PRP treatment compared to baseline with the greatest improved noted in smell discrimination at 3-months post-treatment (ΔD: 2.82, 1.76-3.87, p<0.0001). In contrast, placebo intervention resulted in an improvement in smell threshold at 3-months (ΔT: 1.75, 0.41-3.09, p=0.011) but no changes in the other components of olfaction.

Table 2:

Change of olfactory score from baseline to 1-month and 3-month post-intervention visit

		Month 1 vs. Baseline				Month 3 vs. Baseline
	Arm	Change	Lower CI	Upper CI	p-value	Change	Lower CI	Upper CI	p-value
Total TDI
	Placebo	1.17	−1.99	4.32	0.464	2.58	−0.13	5.29	0.061
	Treatment	4.31	1.69	6.93	0.002	6.25	3.85	8.65	<.0001
	Difference	3.15	−0.96	7.25	0.131	3.67	0.05	7.29	0.047
T Score
	Placebo	0.42	−1.07	1.91	0.579	1.75	0.41	3.09	0.011
	Treatment	2.04	0.79	3.30	0.002	1.82	0.64	3.00	0.003
	Difference	1.63	−0.32	3.57	0.100	0.07	−1.71	1.85	0.935
D score
	Placebo	0.92	−0.36	2.19	0.157	0.42	−0.78	1.62	0.488
	Treatment	1.14	0.06	2.21	0.038	2.82	1.76	3.87	<.0001
	Difference	0.22	−1.45	1.89	0.793	2.40	0.80	4.00	0.004
I score
	Placebo	−0.17	−1.72	1.39	0.831	0.42	−0.99	1.83	0.555
	Treatment	1.54	0.23	2.85	0.022	1.53	0.29	2.78	0.017
	Difference	1.71	−0.32	3.74	0.098	1.12	−0.76	3.00	0.239
VAS
	Placebo	1.20	0.05	2.35	0.040	1.25	0.27	2.23	0.014
	Treatment	1.50	0.51	2.49	0.004	2.13	1.33	2.93	<.0001
	Difference	0.30	−1.22	1.82	0.694	0.88	−0.38	2.15	0.167

TDI - Threshold Discrimination Identification Score (Sniffin’ Sticks)

CI - 95% Confidence Interval

When assessing subjective changes in smell function, both the placebo and the PRP arms demonstrated a significant improvement in VAS at 1-month and 3-months compared to baseline (Table 2). VAS scores improved Δ1.2, 0.05-2.35, p=0.04 (1-month) and Δ1.25, 0.27-2.23, p=0.014 (3-month) in the placebo arm and Δ1.5, 0.51-2.49, p=0.004 (1-month) and Δ2.13, 1.33-2.93, p<0.0001 (3-month) in the PRP arm.

PRP treatment resulted in a 3.67 point greater improvement in olfaction (TDI score) compared to the placebo group at 3-months (95% CI: 0.05-7.29, p=0.047) in a mixed linear model adjusted for baseline olfactory score (Table 2, Figure 2). There was also a 2.40 point greater improvement in discrimination scores in the PRP versus placebo group at 3-months (95% 0.80-4.00, p=0.004). There was no statistical difference in the improvement of overall olfaction score or individual component scores between PRP and placebo at 1-month post-treatment. The change in olfaction threshold and identification were also similar in both study arms at 3-months. No significant difference was found in the change of subjective olfaction scores (VAS) at either month 1 or month 3 between placebo and intervention (Table 2, Figure 2).

Figure 2:

Measured psychophysical (TDI) and subjective (VAS) olfaction scores at baseline, 1 month, and 3 month post-treatment, from linear mixed models adjusted for baseline score.

Error bars represent 95% confidence intervals.

In evaluating responder rate, at 1-month post-intervention, 3 of 12 (25.0%) subjects in the placebo arm had clinically significant improvement in olfactory function compared to 7 of 17 (41.20%) subjects in the PRP arm (OR 2.0, 95% exact bootstrap CI 0.4-17.0). By completion of the trial (3-month post-intervention), the responder rate was 8.3% in the placebo arm (1 of 12) compared to 57.1% (8 of 14) of subjects in the PRP arm (OR 12.5, 95% exact bootstrap CI 2.2-116.7).

None of the participants reported long-standing adverse effects related to the injections. Short-term side effects were related to the injection itself and consisted of nasal congestion and pressure that lasted up to 24 hours, experienced by both PRP and placebo arms. One participant in the placebo arm reported photophobia lasting for a few hours post-injection that self-resolved. Follow up endoscopic visualization showed no gross effects to the olfactory cleft mucosa at 3 months post-treatment.

Discussion

In this single-blinded randomized controlled study, PRP treatment resulted in a greater improvement in overall olfaction scores compared to placebo with a 12.5 times greater likelihood in achieving a treatment response at 3-months. Submucosal injections of PRP into the olfactory cleft were well tolerated without significant adverse effects and did not worsen smell function, as previously noted in our pilot study and other studies utilizing intranasal PRP.^{20–22,28,29} This data suggests that PRP has the potential as a safe treatment option for patients with COVID-19 smell loss.

However, there was no statistical difference in overall subjective improvement between the PRP and placebo arms. Both arms of the study demonstrated significant improvement at 1-month and 3-month post-treatment. The lack of difference may be due to an underpowered study sample that did not account for the magnitude of spontaneous recovery or placebo effect. Furthermore, the greatest improvement with PRP therapy was seen in smell discrimination. Subjective olfactory improvement is likely variable with each individual placing a different weighted importance on smell intensity, discrimination, and identification. However, it has also been shown that subjective improvement lags objective recovery in COVID-19 related OD.³⁰ Thus it is possible that subjective improvement may be more notable with a longer follow-up period.

In their study, Steffens et al reported olfaction outcomes using a cohort of subjects who underwent a single intranasal injection of PRP with a 1 month follow up²² and found that PRP treatment resulted in higher TDI scores compared to control. Our two studies differ in that ours was a randomized, blinded study that involved a placebo injection, had a longer follow-up period of 3 months, and included only subjects who had failed olfactory training. Both studies had similar levels of improvement in TDI scores following PRP treatment but with different follow up periods (Δ6.25 points at 3 months versus Δ6.7 point at 1 month, respectively). In our study, the control group had greater olfactory improvement (Δ3.0 points at 3 months versus Δ0.5 points at 1 month). This difference in olfactory improvement between the two studies’ control groups likely reflects the placebo effect of receiving a sham procedural intervention and the differences in spontaneous resolution with a longer follow-up period.

Although not a named outcome of this study, we did make a note of those with coinciding parosmia as many COVID-19 patients with smell loss also suffer from smell distortion. We did not notice any change in parosmia following PRP treatment. Additionally, the presence of parosmia did not affect objective olfaction recovery based on adjusted linear mixed models (data not shown). While our analysis controlled for baseline olfactory scores, we also noted that duration of OD did not affect smell recovery. The study recruited subjects over the course of a year (2021-2022) with a least 6 months of OD, and while there have been multiple variants of SARS-CoV2 during this period, the randomization between PRP and placebo was well balanced over the entire duration of enrollment.

Limitations of this study include the small sample size of this study. PRP treatments resulted in significantly improved olfactory function compared to placebo with a higher responder rate, but the wide confidence intervals in our model highlights the variability of response and small sample size and thus the high odds ratio should be interpreted with caution. Two participants in the placebo arm were responders at month 1 (Δ6.0 TDI points) but were no longer responders at month 3 (Δ4.0 points at month 3). This difference is likely within the anticipated retesting margin of error. Future larger studies will allow for a better understanding of effect size between PRP and placebo. In performing a power analysis based off our pilot study, we estimated that the ability to detect a type I error with 80% power (α=0.05), would require 20 patients (10 control, 10 experimental). However, this analysis did not account for olfaction improvement in the placebo arm which is likely due to spontaneous recovery, a placebo effect of obtaining an intervention, or the effects of other ongoing, pre-trial treatments. The effect sizes from this clinical trial will help guide sample size calculations for future studies.

Other limitations include the lack of prior data to inform the optimal dosage or concentration of our PRP injections, which may have an impact on olfaction recovery. Given our past experience, we injected 1mL of PRP into the olfactory cleft bilaterally at two different sites (each 0.5ml) along the superoposterior septum, a region previously shown to have high concentrations of olfactory nerve fibers.³¹ Steffens et al utilized our protocol and injection volume in their recent PRP study.²² In this study, our PRP preparation technique resulted in an average 5.9-fold increase in platelet concentration compared to whole blood (Supplemental Fig 2). This yield is in keeping with prior clinical studies for PRP preparation³² though further studies are required to determine the optimal PRP therapy protocols for OD. Similarly, a better understanding of the mechanism of action in the use of PRP for post-viral olfactory loss is warranted and would benefit from preclinical studies.

Conclusion

In this randomized controlled trial, treatment with intranasal PRP resulted in a greater improvement in measured olfactory function compared to placebo for COVID-19 related OD. Yet there was no subjective olfactory improvement over placebo. Given the paucity of definitive therapeutic options for post-viral OD, PRP therapy may be a promising addition to existing therapies such as olfactory training and steroid irrigations. However, it would be important to counsel potential patients that subjective improvement following PRP therapy can vary by individual. Larger studies are required to better determine optimal candidacy, further assess efficacy, and standardize protocols.

Supplementary Material

supinfo

Supplemental Figure 1: Endoscopic view of the left olfactory cleft with the treatment sites depicted by asterisks (*). 1ml PRP or sterile saline was injected submucosally at two sites along the superoposterior septum within the olfactory cleft, starting approximately 1cm posterior to the head of the MT. The same procedure was performed bilaterally. S=Septum, MT=middle turbinate

Supplemental Figure 2: The whole blood and isolated PRP from 9 subjects in the PRP arm underwent complete blood count and total platelet count (100/mm³) analysis to determine the change in platelet concentration following PRP preparation. Compared to their respective whole blood, PRP samples resulted in an average 5.9-fold increase in platelet concentration.