pysradb is excellent for this. It's a command-line tool (also has a Python API) that queries SRA/GEO metadata directly. Install is just:

bash
pip install pysradb

Step 1 — Search for COVID-19 bulk RNA-seq studies:

bash
pysradb search "COVID-19 RNA-Seq bulk" --db-search --max-results 50

or search for the SARS-CoV-2 human host studies specifically:

bash
pysradb search "SARS-CoV-2 human RNA-Seq" --max-results 100 > covid_rnaseq_hits.txt

Step 2 — Once you have a GSE accession (e.g. from GEO or the search above), get its SRP:

bash
pysradb gse-to-srp GSE152418

Step 3 — Fetch full metadata including patient attributes (age, sex, disease status):

bash
pysradb metadata SRP268271 --detailed

The --detailed flag is the key one for your use case — it pulls the BioSample attributes where submitters deposit clinical/patient metadata like host_sex, host_age, host_disease, etc. These fields vary by submission, but many COVID-19 patient datasets include them.

Step 4 — Save to a TSV for easy filtering:

bash
pysradb metadata SRP268271 --detailed --saveto metadata.tsv

Then open in Excel/R/pandas and filter for library_strategy == RNA-Seq, check for age/sex columns, etc.

Important caveat (as noted above by GenoMax): patient-level clinical data like exact age is sometimes protected or simply not deposited. What you'll reliably find is what the submitter chose to make public in BioSample. Many COVID-19 cohort studies do include age group, sex, and severity — but it varies by study. The --detailed output will show you exactly what each study provides before you commit to downloading anything.