Mortality in ME/CFS: an updated analysis of memorial records

The question of whether ME/CFS contributes to premature death is one that the patient and advocacy community has worked hard to document, often because the medical literature hasn't kept pace with the lived experience. The National CFIDS Foundation's online memorial list — a voluntary archive of deaths in the ME/CFS community — represents one such effort. A 2025 paper by Sirotiak and Amro in Fatigue: Biomedicine, Health & Behavior analyzed the entries in this list and presented descriptive statistics on age at death, illness duration, and reported cause of death. The paper is a legitimate attempt to document a real and under-documented phenomenon. It is also methodologically limited in ways that matter for interpreting what it says — and what it cannot say — about ME/CFS mortality.

What the Memorial List Is and How It Was Analyzed

The National CFIDS Foundation maintains an online memorial list for the ME/CFS community — a place where family members and loved ones can submit entries honoring people who have died, whom they believe died with or from ME/CFS. The list is voluntary and open to submission. It is maintained as a community memorial resource, not as a systematic epidemiological database.

Sirotiak and Amro coded the entries in this list, extracting information about reported age at death, the cause of death as the submitter described it, and the duration of illness before death. They then calculated descriptive statistics: average age at death across the sample, distribution of reported causes, and average illness duration. The paper presents these statistics as an analysis of mortality in ME/CFS.

At the level of description — here is what entries on this memorial list say about the people they commemorate — the analysis does what it attempts. The descriptive statistics are accurate summaries of the entries that were submitted. The methodological questions arise when those statistics are interpreted as characteristics of ME/CFS mortality as a whole.

The Methodological Limitations: No Denominator, No Control Group

The central methodological problem with analyzing a voluntary memorial list as an epidemiological data source is the absence of a denominator. An epidemiological study of mortality needs to know not just how many people died and what they died of, but how many people in the relevant population were alive — the denominator from which the mortality rate is calculated. Without a denominator, mortality rates cannot be computed. You cannot calculate whether ME/CFS patients are dying earlier than expected without knowing how many ME/CFS patients there are, how old they are, and what mortality rates would be expected in a comparable population without ME/CFS.

The National CFIDS Foundation memorial list has no denominator. It contains entries submitted by people who chose to submit them. It does not contain entries for ME/CFS patients who died and whose families did not know the list existed, did not choose to submit an entry, or did not associate the death with ME/CFS. It does not contain entries for ME/CFS patients who are still alive. The subset of deaths represented in the memorial list is determined by who submitted entries, not by the actual distribution of deaths in the ME/CFS population.

There is also no control group. Epidemiological conclusions about whether ME/CFS patients die younger, or die of different causes, or die at higher rates than the general population require a comparison group — people without ME/CFS matched on age, sex, and other relevant characteristics. Without a comparison group, there is no baseline against which to evaluate whether the ages at death or cause-of-death distribution in the memorial list are different from what would be expected.

Selection Bias: Who Gets Submitted to Memorial Lists

Voluntary memorial lists are subject to selection bias — the characteristics of people who get submitted are systematically different from the characteristics of the broader population. Families who know about the National CFIDS Foundation memorial list, who are engaged enough with the ME/CFS advocacy community to submit an entry, and who attribute the death to ME/CFS are likely different from the broader population of people with ME/CFS whose deaths may never be associated with the condition at all.

This matters for interpreting the descriptive statistics. If sicker patients, or patients who had more severe long-term illness, are more likely to have families engaged with ME/CFS advocacy communities, then the memorial list may over-represent the severe end of the disease spectrum. If deaths from certain causes — suicide, organ failure from severe malnutrition — are more likely to be attributed to ME/CFS by submitting families than deaths from incidental causes like cancer or accidents, then the cause-of-death distribution in the list does not reflect the actual cause-of-death distribution in ME/CFS patients broadly.

None of this means the entries in the memorial list are fabricated or that the deaths described are not genuine. It means the sample is not a random or representative sample of ME/CFS mortality, and cannot be treated as one for epidemiological purposes.

What the Paper Can and Cannot Tell Us

What the Sirotiak and Amro analysis can tell us is what the entries on this specific memorial list look like: the average age at death reported by submitters, the causes of death submitters attributed, the durations of illness submitters reported. This is descriptive documentation of a community memorial archive. It has value as a record of community experience and as qualitative evidence of the human cost of ME/CFS — real people, real deaths, real illnesses that families are grieving.

What it cannot tell us, and what its statistical structure cannot support, is the rate at which ME/CFS patients die, whether that rate differs from the general population, what causes of death are over- or under-represented in ME/CFS compared to a comparable population, or whether ME/CFS reduces life expectancy. These questions require population-based denominators and comparison groups that the memorial list dataset does not have.

Reading This Alongside the Population-Based Evidence

The Sirotiak and Amro paper should be read in the context of the studies that do have population-based methodology. Roberts and colleagues' 2016 Lancet study used UK Clinical Practice Research Datalink data with over 2,000 ME/CFS patients matched to general population controls. It found no significant elevation in all-cause mortality, and found significantly elevated suicide risk. The ME Association's 2023 mortality review synthesized available population data alongside ONS death certificate records and reached the honest conclusion that no firm conclusions about life expectancy can currently be drawn.

The three sources together — Roberts 2016, ME Association 2023, and Sirotiak and Amro 2025 — represent the current state of ME/CFS mortality knowledge at different levels of methodological rigor and different types of data sources. They converge on the finding that suicide is the most consistently elevated mortality signal. They diverge in what they can say about all-cause mortality and life expectancy, and that divergence reflects genuine uncertainty rather than contradiction.

For the reader who wants to understand what is actually known about ME/CFS mortality, the honest answer is that the population-based evidence does not find dramatically elevated all-cause mortality, that elevated suicide risk is the clearest signal, and that the voluntary memorial list data document real human cost without providing the epidemiological infrastructure to calculate rates or draw comparative conclusions. The people in the memorial list are real. The statistics that can be derived from a voluntary, uncontrolled archive are not sufficient to answer the questions that patients and families most need answered — and being clear about that distinction is how we maintain the credibility necessary to advocate for research that eventually will.