Abstract
This observational cohort study assesses the occurrence of post–COVID-19 condition symptoms in Medicare enrollees prescribed nirmatrelvir and molnupiravir.
While the COVID-19 pandemic appears to be winding down, its effects are still felt by the millions of people worldwide experiencing post–COVID-19 condition (PCC, or long COVID).1 The antiviral drug nirmatrelvir (marketed as Paxlovid [Pfizer], in combination with ritonavir) and molnupiravir (Lagevrio [Merck]) are recommended as first- and second-line treatments for acute illness in patients with specific risk factors (eg, diabetes).2 However, there are still no US Food and Drug Administration–approved drugs for the treatment or prevention of PCC. Recent studies among US veterans (mostly male) suggest that nirmatrelvir and molnupiravir reduce the risk of some sequelae of COVID-19.3,4 We performed a cohort study of the 2 drugs in PCC in older patients who were Medicare enrollees.
Methods
The cohort came from Medicare enrollees aged 65 years or older diagnosed with COVID-19 between January and September 2022. COVID-19 was identified with an outpatient International Statistical Classification of Diseases, Tenth Revision, Clinical Modification code of U07.1. In January 2022, free home COVID-19 tests became available and not all positive self-tests were captured in Medicare data. Therefore, we also considered the prescription of nirmatrelvir or molnupiravir to be indicative of COVID-19 because no other indications existed. Following previous work,5 we identified PCC based on the World Health Organization (WHO) consensus clinical definition.6 Any new occurrence (not present prior to COVID-19 diagnosis) of the 11 symptoms between 4 to 12 weeks after infection was considered as PCC. We used an extended Cox regression with propensity score adjustment to examine the 2 drugs and the incidence of PCC. We included age, sex, race, geographic region, dual eligibility, low-income subsidy, and 51 chronic comorbidities as covariates as included in the Medicare data (eMethods, eTable in Supplement 1). This study was declared not human participant research by the Office of Human Research Protection at the National Institutes of Health. Statistical analyses were conducted using SAS version 7.15 (SAS Institute Inc) and a 2-sided significance at P < .05. This study followed the STROBE reporting guideline.
Results
Overall, among 3 975 690 outpatients with COVID-19, 57% remained in our study after exclusion. Among them, 19.5% received nirmatrelvir and 2.6% received molnupiravir. PCC incidence among patients receiving nirmatrelvir was 11.8%, 13.7% for molnupiravir, and 14.5% for neither, absolute risk reduction was 2.7% for nirmatrelvir, 0.8% for molnupiravir, with hazard ratios (HRs) of 0.87 (95% CI, 0.86-0.88; P < .001) for nirmatrelvir and 0.92 (95% CI, 0.90-0.94; P < .001) for molnupiravir, compared with no treatment (Table 1). Sensitivity analysis of only patients with the COVID-19 code showed a similar pattern but smaller effect sizes (nirmatrelvir: HR, 0.93 [95% CI, 0.92-0.94; P < .001], molnupiravir: HR, 0.96 [95% CI, 0.93-0.99; P = .001]). In an interaction analysis, we found significantly smaller effect sizes in females than males (HRs for nirmatrelvir: 0.89 vs 0.84; molnupiravir: 0.95 vs 0.88). Female sex; Asian, Black, and Hispanic races; and indicators of low income were associated with increased risk of PCC. The most common symptoms in PCC were fatigue (29.9%), dyspnea (22.4%), and cough (21%) (Table 2).
Table 1. Hazard Ratio Based on Cox Regression Modela.
| Index variable | Reference | No. (%) | Hazard ratio (95% CI) | Event rate, % (95% CI) | ||
|---|---|---|---|---|---|---|
| Index group | Reference group | Absolute risk reductionb | ||||
| Nirmatrelvir | None | 439 134 (19.5) | 0.87 (0.86 to 0.88) | 11.8 (11.7 to 11.9) | 14.5 (14.4 to 14.6) | 2.7 |
| Molnupiravir | None | 58 914 (2.6) | 0.92 (0.90 to 0.94) | 13.7 (13.5 to 14.0) | 14.5 (14.4 to 14.6) | 0.8 |
| Female | Male | 1 313 415 (58.5) | 1.17 (1.16 to 1.18) | 14.5 (14.4 to 14.6) | 13.2 (13.1 to 13.2) | −1.3 |
| Age, y | ||||||
| 70-74 | 65-69 | 656 324 (29.2) | 0.78 (0.77 to 0.79) | 12.7 (12.7 to 12.8) | 12.0 (11.9 to 12.1) | −0.7 |
| 75-79 | 65-69 | 509 291 (22.7) | 0.70 (0.69 to 0.71) | 14.2 (14.1 to 14.3) | 12.0 (11.9 to 12.1) | −2.2 |
| 80-84 | 65-69 | 324 008 (14.4) | 0.64 (0.63 to 0.66) | 15.8 (15.7 to 16.0) | 12.0 (11.9 to 12.1) | −3.8 |
| ≥85 | 65-69 | 313 754 (14.0) | 0.61 (0.60 to 0.63) | 16.9 (16.7 to 17.0) | 12.0 (11.9 to 12.1) | −4.9 |
| Racec | ||||||
| Asian | White | 81 073 (3.6) | 1.10 (1.07 to 1.12) | 13.3 (13.0 to 13.5) | 13.9 (13.9 to 14.0) | 0.6 |
| Black | White | 82 249 (3.7) | 1.24 (1.22 to 1.27) | 15.3 (15.0 to 15.5) | 13.9 (13.9 to 14.0) | −1.4 |
| Hispanic | White | 93 325 (4.2) | 1.02 (1.00 to 1.04) | 15.4 (15.1 to 15.6) | 13.9 (13.9 to 14.0) | −1.5 |
| Otherd | White | 93 011 (4.1) | 1.04 (1.02 to 1.06) | 12.4 (12.1 to 12.6) | 13.9 (13.9 to 14.0) | 1.5 |
| Income | ||||||
| Dual eligibility | Nondual | 244 874 (10.9) | 1.06 (1.05 to 1.08) | 16.6 (16.5 to 16.8) | 13.6 (13.5 to 13.6) | −3.0 |
| Low-income subsidy | Nondual | 21 049 (0.9) | 1.07 (1.03 to 1.10) | 16.4 (15.9 to 16.9) | 13.6 (13.5 to 13.6) | −2.8 |
| Region | ||||||
| Midwest | Northeast | 443 777 (19.8) | 0.91 (0.90 to 0.92) | 13.3 (13.2 to 13.4) | 14.1 (14.1 to 14.2) | 0.8 |
| South | Northeast | 869 144 (38.7) | 0.95 (0.94 to 0.97) | 14.2 (14.2 to 14.3) | 14.1 (14.1 to 14.2) | −0.1 |
| West | Northeast | 423 314 (18.8) | 1.04 (1.03 to 1.06) | 13.7 (13.6 to 13.8) | 14.1 (14.1 to 14.2) | 0.4 |
| Other | Northeast | 16 706 (0.7) | 0.52 (0.49 to 0.56) | 14.0 (13.5 to 14.5) | 14.1 (14.1 to 14.2) | 0.1 |
The 51 chronic comorbidities that are included as covariates are not shown in this table.
Absolute risk reduction is the difference of raw event rate between reference and index groups (reference – index). It is possible that the directionality of absolute risk reduction can be different from that indicated by the hazard ratio adjusted for covariates and propensity scores.
Race categories are included as found in Medicare data and are included as potential factors in the outcomes of COVID-19.
The race classification aligns with that in the US Centers for Medicare & Medicaid Services Virtual research data Center database, which was American Indian or Alaska Native, Asian, Black, Hispanic, White, other, and unknown. We combined American Indian or Alaska Native, other, and unknown into other because of the small numbers in these categories.
Table 2. Characteristics of Patients With Post–COVID-19 Condition.
| Characteristic | No. of patients (%) | Standardized mean difference | ||||
|---|---|---|---|---|---|---|
| Overall (N = 313 262) | Nirmatrelvir (n = 51 658) | Molnupiravir (n = 8089) | None (n = 253 617) | Nirmatrelvir vs none | Molnupiravir vs none | |
| Post–COVID-19 condition symptom | ||||||
| Fatigue/malaise/weakness | 93 653 (29.9) | 15 049 (29.1) | 2292 (28.3) | 76 338 (30.1) | −0.02 | −0.04 |
| Dyspnea | 70 306 (22.4) | 10 810 (20.9) | 1811 (22.4) | 57 707 (22.8) | −0.04 | −0.01 |
| Cough | 65 660 (21.0) | 10 910 (21.1) | 1737 (21.5) | 53 035 (20.9) | 0.01 | 0.01 |
| Chest pain | 55 506 (17.7) | 8905 (17.2) | 1432 (17.7) | 45 185 (17.8) | −0.02 | 0.00 |
| Palpitations | 37 734 (12.0) | 5943 (11.5) | 919 (11.4) | 30 887 (12.2) | −0.02 | −0.03 |
| Headache | 25 704 (8.2) | 3983 (7.7) | 701 (8.7) | 21 033 (8.3) | −0.02 | 0.01 |
| Muscle/joint pain | 23 174 (7.4) | 4205 (8.1) | 605 (7.5) | 18 368 (7.2) | 0.03 | 0.01 |
| Memory problem | 13 093 (4.2) | 2194 (4.2) | 350 (4.3) | 10 552 (4.2) | 0.00 | 0.01 |
| Cognitive impairment | 8096 (2.6) | 1070 (2.1) | 190 (2.3) | 6837 (2.7) | −0.04 | −0.02 |
| Sleep disturbance | 3844 (1.2) | 699 (1.4) | 86 (1.1) | 3059 (1.2) | 0.01 | −0.01 |
| Loss of taste/smell | 1321 (0.4) | 220 (0.4) | 30 (0.4) | 1071 (0.4) | 0.00 | −0.01 |
| Female | 190 372 (60.8) | 31 176 (60.4) | 4684 (57.9) | 154 571 (60.9) | −0.01 | −0.06 |
| Age range, y | ||||||
| 65-69 | 53 348 (17.0) | 10 264 (19.9) | 1226 (15.2) | 41 873 (16.5) | 0.09 | −0.04 |
| 70-74 | 83 569 (26.7) | 15 800 (30.6) | 2119 (26.2) | 65 674 (25.9) | 0.11 | 0.01 |
| 75-79 | 72 104 (23.0) | 12 166 (23.6) | 1985 (24.5) | 57 978 (22.9) | 0.02 | 0.04 |
| 80-84 | 51 308 (16.4) | 7581 (14.7) | 1436 (17.8) | 42 317 (16.7) | −0.05 | 0.03 |
| ≥85 | 52 933 (16.9) | 5847 (11.3) | 1323 (16.4) | 45 775 (18.0) | −0.18 | −0.04 |
| Racea | ||||||
| Asian | 10 762 (3.4) | 2017 (3.9) | 145 (1.8) | 8607 (3.4) | 0.03 | −0.09 |
| Black | 12 549 (4.0) | 1297 (2.5) | 227 (2.8) | 11 026 (4.3) | −0.09 | −0.07 |
| Hispanic | 14 327 (4.6) | 1719 (3.3) | 284 (3.5) | 12 328 (4.9) | −0.07 | −0.06 |
| White | 264 131 (84.3) | 44 414 (86.0) | 7172 (88.7) | 212 630 (83.8) | 0.06 | 0.13 |
| Otherb | 11 493 (3.7) | 2211 (4.3) | 261 (3.2) | 9026 (3.6) | 0.04 | −0.02 |
| Income | ||||||
| Dual eligibility | 40 725 (13.0) | 3206 (6.2) | 638 (7.9) | 36 892 (14.5) | −0.25 | −0.19 |
| Nondual, low-income subsidy | 3454 (1.1) | 378 (0.7) | 82 (1.0) | 2994 (1.2) | −0.04 | −0.01 |
| Nondual, no low-income subsidy | 269 083 (85.9) | 48 074 (93.1) | 7369 (91.1) | 213 731 (84.3) | 0.25 | 0.19 |
| Region | ||||||
| Northeast | 69 890 (22.3) | 11 205 (21.7) | 1059 (13.1) | 57 639 (22.7) | −0.02 | −0.23 |
| Midwest | 59 226 (18.9) | 9924 (19.2) | 1569 (19.4) | 47 749 (18.8) | 0.01 | 0.02 |
| South | 123 707 (39.5) | 19 899 (38.5) | 4444 (54.9) | 99 416 (39.2) | −0.01 | 0.32 |
| West | 58 099 (18.5) | 10 370 (20.1) | 949 (11.7) | 46 801 (18.5) | 0.04 | −0.18 |
| Other | 2340 (0.7) | 260 (0.5) | 68 (0.8) | 2012 (0.8) | −0.03 | 0.01 |
Race categories are included as found in Medicare data and are included as potential factors in the outcomes of COVID-19.
The race classification aligns with that in the US Centers for Medicare & Medicaid Services Virtual research data Center database, which was American Indian or Alaska Native, Asian, Black, Hispanic, White, other, and unknown. We combined American Indian or Alaska Native, other, and unknown into other because of the small numbers in these categories.
Discussion
Consistent with the findings of Xie et al,3,4 we found that nirmatrelvir and molnupiravir were associated with a small reduction in incidence of PCC. Our effect sizes are smaller than those of Xie et al3,4 (absolute risk reduction, nirmatrelvir 4.5%; molnupiravir 3.0%) but our sample size is 8-fold larger. We also have a more balanced sex ratio (female 59% vs 14%), which is important because PCC is more common in females. The smaller effect sizes in females may explain our overall smaller effect sizes. We used the WHO consensus definition based on symptoms rather than disease diagnosis (eg, ischemic heart disease), which is more akin to how PCC is identified clinically. Limitations of our study include not incorporating vaccination status because of incomplete data, use of prescription of the drugs as evidence of COVID-19, and restriction to patients 65 years or older. The current approved use of the 2 drugs is for the prevention of severe acute COVID-19. Our findings suggest that they may also have a role in preventing PCC.
eMethods
eReferences
eTable 1. List of 51 Comorbidities
Data Sharing Statement
References
- 1.Davis HE, McCorkell L, Vogel JM, Topol EJ. Long COVID: major findings, mechanisms and recommendations. Nat Rev Microbiol. 2023;21(3):133-146. doi: 10.1038/s41579-022-00846-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hammond J, Leister-Tebbe H, Gardner A, et al. ; EPIC-HR Investigators . Oral nirmatrelvir for high-risk, nonhospitalized adults with COVID-19. N Engl J Med. 2022;386(15):1397-1408. doi: 10.1056/NEJMoa2118542 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Xie Y, Choi T, Al-Aly Z. Association of treatment with nirmatrelvir and the risk of post-COVID-19 condition. JAMA Intern Med. 2023;183(6):554-564. doi: 10.1001/jamainternmed.2023.0743 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Xie Y, Choi T, Al-Aly Z. Molnupiravir and risk of post-acute sequelae of covid-19: cohort study. BMJ. 2023;381:e074572. doi: 10.1136/bmj-2022-074572 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Fung KW, Baye F, Baik SH, Zheng Z, McDonald CJ. Prevalence and characteristics of long COVID in elderly patients: An observational cohort study of over 2 million adults in the US. PLoS Med. 2023;20(4):e1004194. doi: 10.1371/journal.pmed.1004194 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.World Health Organization . A clinical case definition of post COVID-19 condition by a Delphi consensus. October 6, 2021. Accessed August 12, 2023. https://www.who.int/publications/i/item/WHO-2019-nCoV-Post_COVID-19_condition-Clinical_case_definition-2021.1
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods
eReferences
eTable 1. List of 51 Comorbidities
Data Sharing Statement