Currently, there is no validated diagnostic biomarker for ME/CFS, and the disease is diagnosed based on clinical criteria such as the 2021 NICE guidelines, or the Canadian Consensus Criteria.
The bias and stigma associated with the ME/CFS, alongside limited awareness, gaps in medical knowledge, and a high symptom overlap with other illnesses, leads to delays in accurate diagnosis.
In fact, one study, published in 2021, made the following observation:
“Many people with ME/CFS are experiencing a serious and potentially harmful delay in having their diagnosis confirmed. Many individuals are diagnosed within the first 2 years, but a considerable number are diagnosed after more than 10 years.”
As such, identifying ways to improve the ME/CFS diagnostic process is essential, including obtaining a clearer picture of the steps many people go through before the disease is diagnosed.
Therefore, to try and understand which diagnoses might be given before one of ME/CFS in children and young people aged between 6 and 27 years old, a team of researchers in Germany, including Uta Behrends, looked at health records from a large medical insurance company; the company, Techniker Krankenkasse, is thought to cover about 15% of the German population.
Importantly, the researchers stated their aim was to:
“Identify symptoms, clinical conditions, and patterns of healthcare utilisation in form of ICD-10-GM codes from the year preceding an ME/CFS diagnosis (G93.3) that could serve as early diagnostic indicators of ME/CFS for clinicians and health insurances”
In Germany, the coding system used to record diagnoses made in clinical practice is the International Classification of Diseases 10th Revision German Edition (ICD-10-GM). In this study, the code G93.3 is used to identify cases of ME/CFS. While this code is not specific to ME/CFS – it is used to classify post-viral and related fatigue syndromes, which are characterised by persistent fatigue following a viral infection – it is thought to be the code which “best reflects ME/CFS symptoms”, as noted by Gemma Samms and Prof. Chris Ponting in their 2025 study estimating the prevalence of ME/CFS.
To assess the medical records, the team used a study design known as a ‘matched case-control’. Here each of the 6,077 cases – people with ME/CFS – were matched with 5 healthy controls, who were the same sex, born in the same year, and lived in the same area. In theory, using this design means that the study accounts for the potential effects of sex, age, and location.
Most cases were aged between 18 and 24 years old and, as in many ME/CFS studies, female. This is unsurprising given that the disease appears to be more common in women compared with men at approximately a 4:1 ratio.
The results showed that there were 44 ICD-10-GM codes associated with an increased chance of going on to receive a G93.3 diagnosis. Interestingly, these codes spanned 13 different diagnostic areas, but most appeared in:
- Mental/behavioural disorders.
- Respiratory diseases.
- Musculoskeletal disorders.
Frequently diagnosed conditions included fatigue, depression, pain disorders, and medically unexplained physical symptoms (termed ‘somatoform disorders’ in this study).
There were also some medical codes, such as those for fibromyalgia, and for ‘mild cognitive impairment’, that, while less frequently diagnosed, were strongly associated with an increased chance of going on to receive a G93.3 diagnosis.
Importantly, history of laboratory testing was associated with an increased chance of receiving a G93.3 diagnosis, while a history of vaccination against infection, a significantly lower chance.
Interestingly, four COVID-19 or vaccination-related code classes were identified. Of these codes, a record of ‘post-COVID-19 condition’ had the highest chance of going on to receive an ‘ME/CFS’ diagnosis.
In their paper, the research team explain that studies following people up over time should clarify timing relative to ME/CFS onset, and distinguish between pre-existing conditions, comorbidities, early manifestations, or misdiagnoses.
It is also important to note that this study used the code G93.3 rather than confirming a diagnosis of ME/CFS; it is possible that not all those with a G93.3 diagnosis would meet more stringent ME/CFS criteria. Additionally, medical records were only looked at for the year prior to ME/CFS diagnosis and, as mentioned above, the diagnostic process can commonly take much longer than that meaning that a significant amount of potentially relevant information was not considered.
