The 2-Day CPET Reliability and Validity Debate
The 2-day cardiopulmonary exercise test (CPET) is the most objective biomarker of PEM yet identified — and also the most debated methodological instrument in ME/CFS research. The test measures VO2peak, workload at ventilatory threshold, and other cardiopulmonary parameters on two consecutive days. In healthy people, these parameters are stable or improve slightly on Day 2 (learning effect). In ME/CFS patients, they decline — sometimes substantially — on Day 2, providing an objective, quantitative measure of post-exertional decline (Snell et al. 2013).
1 The Signal
The 2-day CPET signal has been replicated across multiple studies from different research groups:
- Snell et al. (2013): Day-2 VO2peak decline of 13.8% in ME/CFS vs 4.7% in controls (Snell et al. 2013)
- Keller et al. (2024): Day-2 VO2peak decline of 6.3% in ME/CFS; workload@VT decline of 19% (Keller et al. 2024)
- van Campen et al. (2020): Severe ME/CFS showed 21.7% VO2peak decline vs 10.4% for mild — CPET discriminates severity Campen, Rowe, and Visser (2020)
- Lim et al. (2020): Meta-analysis of 5 studies — large effect sizes for VO2peak (g=1.11) and workload@VT (g=1.13) (Lim et al. 2020)
A case report of identical twins discordant for ME/CFS — where genetics, upbringing, and baseline fitness were controlled — showed a 13% VO2peak decline only in the affected twin (Giloteaux, Hanson, and Keller 2016), providing within-subject replication of the Day-2 decline.
2 The Limitations
Despite the replicated signal, several methodological concerns qualify the 2-day CPET’s status as a gold standard:
Network dependency: All positive 2-day CPET studies originate from a small network of allied research groups — the Workwell Foundation (Snell, Stevens), the van Campen/Visser group, and the Keller/Hanson group. These groups share protocols, collaborate, and cite each other. No fully independent replication by a skeptical laboratory with no connection to these networks exists. This is a standard concern in meta-science: when all positive findings come from a single research network, the finding may reflect shared methodology rather than shared biology.
Data integrity concerns: An independent analysis published online identified three specific methodological problems in the Workwell CPET dataset: conflicting data values for the same patients across publications, improbable zero-difference clustering between Day 1 and Day 2 values, and circular reasoning in normal range calculations (ME/CFS Science 2024). These concerns are methodological, not substantive, and have not been formally published or refuted — but they exist and they come from a source outside the CPET network.
Long COVID null result: A 2025 study found no significant Day-1 vs Day-2 VO2peak difference in Long COVID patients with PEM (Gattoni et al. 2025). If Long COVID + PEM patients don’t show the 2-day CPET signal, this complicates the claim that 2-day CPET is specifically measuring PEM across post-infectious syndromes.
Clinical risk: The 2-day CPET can trigger severe and prolonged PEM episodes. It cannot be administered to severe patients (contraindicated), and even in moderate patients carries risk. Standardizing ramp protocols, environmental conditions, and test timing is essential but not always achieved (Van Ness, Snell, and Stevens 2013) (Stevens et al. 2018).
Consequence: The 2-day CPET is the best available objective measure of PEM, but it is imperfect. It should be treated as a strong signal that requires independent replication by skeptical laboratories — not as a settled gold standard. The test carries clinical risk and should only be used when the research question genuinely requires it. For clinical diagnosis, it is not appropriate for most patients.
3 The Broader Question: What Would Replicate the 2-day CPET Signal Independently?
If an independent laboratory (no Workwell/van Campen-Visser/Keller network affiliation, no pre-existing commitment to the 2-day CPET model) were to replicate the Day-2 VO2peak decline in a well-characterized ICC cohort with matched sedentary controls, this would substantially strengthen confidence. Until then, the 2-day CPET remains — like most biomarker findings in ME/CFS — a replicated signal from a single research network.
Consequence: Readers should understand that the 2-day CPET provides objective evidence of post-exertional decline but is not independently validated. The most important contribution of the CPET literature may be its demonstration that PEM has an objective physiological correlate, which refutes claims that PEM is a psychological or behavioral phenomenon.