The multiple sclerosis prodrome

Increased health-care usage.

Population-based studies have emerged in the past 5 years^{4–6,9–13,19–23} and suggest a trend of deteriorating health before clinical onset of MS. For example, in one study conducted in Canada, analysis of health administrative data linked with MS-specific clinical data from >14,000 people with MS and >72,000 controls from the general population matched for age, sex, region and socioeconomic status revealed that the use of health-care services by patients with MS was higher in the 5 years leading up to their first clinical demyelinating event (based on physician-derived diagnostic codes in administrative data) or MS symptom onset (determined by an MS neurologist) than that of controls. In the year before the first clinical demyelinating event, hospitalizations and physician visits were 78% and 88% higher, respectively, for people with MS than for matched controls. Similarly, dispensed prescription medications were 49% higher among patients who went on to develop MS⁶. In a subsequent study conducted in the UK, analysis of primary care data from >10,000 people with MS and nearly 40,000 matched controls revealed more visits to general practitioners among those with MS up to 10 years before their MS was diagnosed (based on codes for MS or CIS recorded in general practitioner records)¹⁹.

The observed higher rates of health-care usage resulted from a myriad of issues. Use of mental health services by people who developed MS was high^6,9,11 — visits to psychiatrists in the 5 years before MS symptom onset were 50% higher than among controls⁹. Fatigue, pain, fibromyalgia, migraine, bowel issues, bladder issues, sleep disturbances and anaemia were also more common among people who developed MS than among controls in the years leading up to MS symptom onset^9,11,19. Women with MS were also less likely to have become pregnant and were more likely to fill a prescription for a hormonal preparation (such as an oral contraceptive) than controls in the 5 years before onset of typical MS symptoms, observations that suggest a behavioural change before the clinical onset of MS⁹. When faced with a range of health complaints as part of the MS prodrome, women may have opted — whether consciously or subconsciously — to delay pregnancy⁹. Though fewer pregnancies might lead to a greater risk of a first clinical demyelinating event²⁴, the potential for reverse causation must be carefully considered.

Cognitive impairment.

Low cognitive performance before onset of typical MS symptoms has been reported in a nested case–control study of >20,000 Norwegian men who entered the mandatory national military service at the age of 18–19 years²⁰. Among the 924 individuals diagnosed with MS, cognitive performance was lower than among the 19,530 healthy controls up to 2 years before MS onset. Among the <50 individuals who developed primary progressive MS (rather than relapsing–remitting), cognitive performance was lower than among the controls up to 20 years before clinical onset of MS²⁰. However, whether these observations are generalizable to the wider population, including women, remains undetermined.

Age and sex effects.

Sex and age seem to affect features of the MS prodrome, although these and other population demographics, such as ethnicity, are underexplored^10,11. In a study conducted in Canada that included 6,863 people with MS and 31,865 matched controls, the odds of anaemia were higher among men with MS than among men in the general population (OR 2.40, 95% CI 1.68–4.29), whereas the increase in odds for women with MS were more modest (OR 1.23, 95% CI 1.04–1.45). Pain was more likely in older adults than in younger adults in the 5 years before a first demyelinating event — the odds ratio increased from 1.76 at <30 years of age to 2.14 at 30–50 years of age and to 2.35 at ≥50 years of age¹¹. All odds ratios were adjusted for socioeconomic status, index year, age and sex. Further work is warranted to delineate the impact of age, sex and other patient demographics on features of the MS prodrome.

Disease course and dermatological complaints.

Another study conducted in Canada identified visits to dermatologists as a prodromal feature of MS that might differentiate the two main disease courses. Visits were more common among those who developed relapsing–remitting MS (n = 1,887) than among those who developed primary progressive MS (n = 171) in the 5 years before onset of typical MS symptoms. These skin-related issues could conceivably reflect an early inflammatory manifestation of MS¹⁰.

Other features.

Further possible features of the MS prodrome have been identified in a systematic review of MS morbidities. The review included 29 studies published before February 2019; in these studies, the period before MS diagnosis or clinical MS onset was examined¹³. Broadly, the review identified consistent evidence that anxiety, depression and migraine as well as a low cognitive performance are all part of the MS prodrome¹³.

Radiologically isolated syndrome.

MRI findings that are characteristic of MS are sometimes detected incidentally in individuals with no typical symptoms of MS but who underwent MRI for other reasons (for example, head trauma). These findings are referred to as the radiologically isolated syndrome (RIS), a formal definition of which was proposed in 2009 (REF.²⁵). In the diagnostic criteria for MS, RIS is distinguished by a lack of typical MS symptoms. RIS was first described in adults but has also been reported in children²⁶.

The exact prevalence of RIS is unknown. One large, hospital-based study in Sweden indicated a prevalence of 0.05% (0.15% among those aged 15–40 years) among 2,105 individuals who underwent MRI for any reason during a 1-year period²⁷. A meta-analysis that included 15,559 individuals with no history of neurological symptoms determined that 0.06% had MRI findings that were suggestive of demyelination²⁸.

Whether all individuals with MS develop abnormalities that are detectable on MRI before they develop typical MS symptoms is unknown. However, studies of individuals with a first clinical demyelinating event have revealed multiple lesions on MRI at first clinical presentation^29–32, suggesting that asymptomatic brain lesions predate the first clinical demyelinating event in a substantial proportion of individuals. Carefully designed prospective MRI studies of individuals prior to the onset of typical MS symptoms are therefore warranted to establish whether RIS overlaps with the prodromal phase and, if so, how.

Furthermore, RIS could itself be a marker of the MS prodrome. This possibility is suggested by the finding that the majority of individuals with RIS have a variety of subtle or non-specific symptoms, most commonly headache^25,33. This observation raises the possibility that some people have clinical symptoms that are part of the prodromal phase of MS.

Regardless of whether RIS is a biomarker of the prodrome, individuals with RIS are at substantial risk of a subsequent clinical demyelinating event and a diagnosis of either relapsing–remitting MS or primary progressive MS³⁴. Consequently, markers that are associated with later development of MS in individuals with RIS are of interest in the context of the MS prodrome, as they could also be markers of prodromal MS. Several studies have been conducted to identify such markers. One large, multicentre, retrospective study included 451 individuals from the USA and Europe who fulfilled the 2009 criteria for RIS²⁵. Of these individuals, 34% developed clinical symptoms of MS within 5 years of their first abnormal MRI, increasing to 51% after 10 years (among 257 participants with 10-year follow-up)^33,35. Factors associated with an increased risk of a clinical demyelinating event were a younger age at development of RIS, male sex, the presence of asymptomatic infratentorial or spinal cord lesions on MRI, and abnormal cerebrospinal fluid (CSF) findings (a high IgG index or the presence of two or more unique oligoclonal bands)^33,35. These factors might also warrant exploration as potential biomarkers of the MS prodrome.

Cognitive impairments have also been reported in individuals with RIS. In a study conducted in France that included people with RIS, MS or neither (26 in each group, matched for age, sex and educational level), individuals with MS or RIS had a lower performance than the controls in a range of cognitive assessments, including the Paced Auditory Serial Addition Test (PASAT), phonemic fluency, the digit-symbol subtest of the Weschler Adult Intelligence Scale, digit span (direct and indirect), cross-tapping and the Go-No-Go test³⁶. Cognitive profiles were generally similar in the RIS and MS groups, although performance was lower among those with MS on the direct digit span and cross-tapping tests. A study conducted in Italy resulted in similar findings — 28% (8 of 29) of individuals with RIS and 35% (9 of 26) of those with MS were classified as having cognitive impairment on the basis of failure in at least two tests. Performance on the PASAT2 and word-list generation were lower among those with MS than among those with RIS³⁷. Similarly, a study conducted in Israel showed that the performance on all cognitive tests was lower than expected (based on age and education) for 30 individuals with RIS who were assessed up to 3 months after diagnosis³⁸. Together, these findings in RIS reinforce the idea that impaired cognition could be a measurable feature of the MS prodrome.

Advanced neuroimaging.

Advanced neuroimaging techniques — beyond those that are used in clinical practice — have been used to assess the characteristics of RIS and the findings could indicate features of the MS prodrome. The results of several studies (mainly case–control studies) of modest size suggest that whole-brain volume and regional brain volumes (for example, thalamic and cerebellar volumes) are lower in individuals with RIS than in healthy controls^39–41. In one study conducted in Italy, normalized whole-brain and cortical volumes were similar in 26 individuals with MS and 19 with RIS but lower in both groups than in 21 healthy controls of similar age and sex³⁷. In this study, lower cortical volumes were associated with poorer performance in a cognitive battery among individuals with RIS.

Advanced MRI also identified cortical lesions in 40% of 15 individuals with RIS⁴² and diffusion tensor imaging revealed microstructural changes in white matter tracts in 18 individuals with RIS compared with 20 healthy controls⁴³. In addition, proton magnetic resonance spectroscopy has shown that N-acetylaspartate and creatine levels are lower in people with RIS than in healthy controls, suggesting abnormalities in brain metabolic pathways⁴⁴. These early studies suggest that advanced neuroimaging could be helpful for identifying biomarkers of the prodromal phase of MS, though further studies are needed to determine their value in this context.

Serum and CSF biomarkers.

In addition to neuroimaging, biomarkers in the blood and/or CSF could be of value in the MS prodrome. Several studies have identified potentially useful markers.

In one nested case–control study, analysis of stored serum samples from 30 US military personnel who went on to develop MS and 30 healthy matched controls showed that serum levels of neurofilament light chain (NfL; a marker of neuro-axonal injury) were higher among individuals who went on to develop MS than among controls at a median of 6 years before the onset of MS symptoms⁴⁵. The results of this small but compelling study, which involved a predominantly non-Hispanic white (60%), male (77%) population, suggest that neuro-axonal injury occurs years before onset of the typical clinical symptoms of MS.

In a study of 75 individuals with RIS, 23 of whom developed CIS during follow-up, high CSF levels of NfL and the presence of unique CSF oligoclonal bands were associated with earlier development of CIS⁴⁶. A high IgG index and the presence of two or more unique oligoclonal bands in the CSF have been reported as risk factors for a clinical demyelinating event in adults and children with RIS^26,33,46,47. These CSF markers could therefore be similarly helpful for the risk stratification of individuals with clinical symptoms of possible prodromal MS even if they do not have RIS. Future studies are needed to address this possibility.

Novel techniques, such as single-cell analyses, could also produce biomarkers of prodromal MS⁴⁸. For example, single-cell RNA sequencing of the CSF from six clinically healthy individuals who were monozygotic twins of patients with MS revealed clonal expansion of CD8⁺ T cells and B cells, indicating early adaptive immune activation⁴⁹. Whether these asymptomatic but high-risk individuals will transition to clinical MS remains to be determined, but the findings illustrate the potential of single-cell analysis to identify very early pathological changes.

Paediatric-onset MS.

The typical clinical symptoms of MS usually emerge in early adulthood but 3–10% of individuals with MS develop neurological symptoms before the age of 16–18 years^50,51. The existence of paediatric-onset MS suggests that the pathobiology of MS begins early in life and that the MS prodrome could be detectable as early as childhood.

A prodrome for paediatric-onset MS has been suggested by the findings of a study conducted in Canada, which involved the analysis of administrative data from 659 children (age <18 years) with MS and 3,294 population controls matched (up to 1:5) for sex, age and region of residence²². The proportion of children who were hospitalized in the year before their first clinical demyelinating event was greater among those with MS than among controls. Similarly, the median number of physician visits was greater among the children with MS over the same period.

In a second study in Canada, administrative data were compared with health-care usage data for mothers with and without a child with MS²¹. In comparison with 624 mothers of children without MS, the 156 carefully matched mothers of children with MS had a higher rate of physician visits for any mood or anxiety disorder up to 5 years before the child’s first demyelinating event. One possible explanation for this observation is that children with MS experienced prodromal symptoms in the years before their first demyelinating event, leading to increased stress and higher mental health service use by their mothers²¹. Taken together, these studies suggest that an MS prodromal phase occurs in paediatric-onset MS, but the clinical features of this prodromal phase are yet to be determined.

As in adults, RIS could be a neuroimaging correlate of the MS prodrome in children. RIS in children seems to be relatively uncommon (for example, only one child had RIS in a population-based study of 3,966 children aged 8–11 years)^25,52 but does seem to be associated with an increased risk of developing clinical symptoms of demyelination when it occurs. In a multicentre international study of 38 children with RIS (defined according to the 2010 criteria for dissemination in space on MRI)⁵³, 42% developed a first clinical event after a median of 5 years on the basis of Kaplan–Meyer analyses and an absolute median of 2 years²⁶. The presence of two or more unique oligoclonal bands in the CSF and asymptomatic spinal cord lesions on MRI were associated with an earlier first clinical event, in agreement with findings in adults, again raising the possibility that these findings are biomarkers of the MS prodrome.

First-degree relatives of people with MS.

Studying first-degree family members of individuals with MS, who are at higher risk of MS than the general population, could yield important insights into prodromal features of MS. In one prospective study of 65 first-degree relatives, vibration sensitivity testing of the distal extremities revealed a poorer performance among participants who were at the highest risk of MS than among those at the lowest risk on the basis of a weighted genetic and environmental risk score⁵⁴. This observation suggests that subtle neurological findings can precede the onset of typical MS symptoms. Among the participants of this study, five (8%) fulfilled the MRI criteria for dissemination in space using the 2010 MS diagnostic criteria without meeting the 2009 Okuda definition of RIS (four of whom were in the high-risk group) and two (3%) fulfilled formal criteria for RIS (one in each of the high-risk and low-risk groups). Similarly, other studies of asymptomatic relatives of individuals with either sporadic or familial MS have shown that ~3–10% have RIS^55,56. Together, the occurrence of RIS and subtle neurological signs in high-risk relatives of patients supports the hypothesis that these characteristics precede clinical symptoms of MS and could therefore be markers of prodromal MS.

Other neurodegenerative diseases.

According to the International Parkinson and Movement Disorder Society, early Parkinson disease (PD) can be divided into three phases: preclinical PD, in which neurodegeneration has commenced but no outward signs or symptoms are apparent; prodromal PD, in which signs and symptoms are present but are insufficient to fulfil the diagnostic criteria for PD; and clinical PD, when a diagnosis of PD can be made on the basis of characteristic motor signs and symptoms^57,58. The concept of a prodromal phase of PD is therefore widely accepted and defining features of this prodrome have been identified, although a precise definition continues to evolve^58,59. Factors that are included in the research criteria for prodromal PD include current age, risk factors for PD (including sex, caffeine use, smoking status, genetic markers and echogenicity of the substantia nigra on ultrasound), clinical non-motor markers (including REM sleep behaviour disorder, olfactory dysfunction and constipation), clinical motor markers (including possible subthreshold parkinsonism on expert examination or abnormal quantitative motor tests), and neuroimaging markers on PET or single-photon emission tomography⁵⁸. These factors enable the calculation of the probability of prodromal PD and are intended to help with identifying individuals for enrolment in clinical trials of neuroprotective agents^57,58.

As in PD, the pathology of Alzheimer disease (AD) is thought to begin years before the onset of typical symptoms, and the disease is thought to progress through a continuum of normal, age-expected changes to mild cognitive impairment and, finally, clinical AD. A conceptual framework and operational criteria have been proposed for preclinical AD for research purposes⁶⁰. These criteria span three stages: biomarker (amyloid-β) detection in stage one, amyloid-β accumulation and markers of neuronal injury in stage two, and subtle cognitive changes in stage three⁶⁰.

The concept that PD and AD progress through a continuum of disease from the initial pathological processes that do not cause typical symptoms to a prodromal phase and then to the development of characteristic signs and symptoms is well established. MS shares some of the neurodegenerative features of these diseases and, although its initial course is more commonly relapsing–remitting than primary progressive, MS is also likely to evolve through preclinical to clinical phases. We can therefore learn from our understanding and definitions of the early stages of PD and AD and apply this knowledge to defining the MS prodrome. In particular, the development of research criteria for preclinical and prodromal MS, including imaging and other biomarkers, could greatly aid the design of studies for the primary and secondary prevention of MS.

Other inflammatory diseases.

Besides neurodegenerative diseases, prodromal phases have been characterized in other relapsing inflammatory conditions, including rheumatoid arthritis, systemic lupus erythematosus (SLE) and inflammatory bowel disease (IBD). For rheumatoid arthritis, six phases of disease have been proposed: (1) the presence of genetic risk factors; (2) the presence of environmental risk factors; (3) systemic autoimmunity (for example, anti-citrullinated protein antibodies and rheumatoid factor); (4) symptoms without arthritis; (5) undifferentiated arthritis; and (6) rheumatoid arthritis⁶¹. People with imaging abnormalities in the absence of arthritis are thought to fit between phases 3 and 4. In a study of samples stored as part of a blood donation programme, at least one autoantibody (rheumatoid factor and/or anti-cyclic citrullinated peptide) was present in nearly half of the 79 individuals who later developed rheumatoid arthritis at a median of 4.5 years before symptom onset⁶². Detection of these autoantibodies in combination with prodromal features and other characteristics (such as family history) has been harnessed to identify individuals who are at high-risk of developing clinical rheumatoid arthritis and offer them enrolment into clinical trials^63–65.

In SLE, autoantibodies, including antibodies to nuclear antigens, have also been detected before the onset of characteristic symptoms, at ~3–8 years before diagnosis^66,67. As in rheumatoid arthritis, there is great interest in whether individuals who are at high risk of SLE might benefit from interventions to delay clinical onset. Similarly, for IBD (Crohn disease and/or ulcerative colitis), several characteristics of the prodromal phase have been suggested. For example, a study of 240,984 young men aged 18–19 years who underwent assessment for military conscription in Sweden showed that an elevated erythrocyte sedimentation rate (≥15 mm/h versus 1 mm/h) was associated with an increased risk of subsequent diagnosis of Crohn disease or ulcerative colitis after a median follow-up time of 33 years⁶⁸. In another study of 1,420 asymptomatic first-degree relatives of individuals with Crohn disease, increased intestinal permeability was associated with an increased risk of a subsequent diagnosis of Crohn disease over a mean follow-up period of 7 years⁶⁹. Taken together, these studies in other relapsing inflammatory diseases suggest that a combination of clinical features and other biomarkers are likely to characterize the MS prodrome and could aid its categorization into distinct phases^23,70.

Shared prodromal features of inflammatory diseases.

Inflammatory, immune-mediated diseases might be expected to share prodromal features, with some evidence suggesting this is the case. In one study conducted in Canada, population-based administrative data from the province of Manitoba on 12,141 individuals with immune-mediated inflammatory disease (IBD, MS or rheumatoid arthritis) and 65,424 controls matched for birth year, sex, region and disease were analysed²³. The incidence of psychiatric disorders was higher among patients with inflammatory immune-mediated diseases as early as 5–10 years before diagnosis. The findings were consistent when the three diseases were considered separately. This study focused on the time period before diagnosis rather than at symptom onset so the findings could be influenced by the variable application of diagnostic criteria or diagnostic delays; however, the 10-year findings are nonetheless likely to reflect features of the prodromal phase²³.

Risk factors versus prodromal features.

As knowledge of the MS prodrome develops, the literature surrounding the causes of MS needs to be re-evaluated as presumed risk factors for MS onset could actually be prodromal features. One salient example is pregnancy and oral contraceptive use. Previously, fewer pregnancies and greater use of oral contraceptives were thought to increase a woman’s risk of developing MS^71–73. However, knowledge of the prodrome suggests reverse causation as a plausible explanation — women in the prodromal phase of MS might instead be changing their behaviour in reaction to their escalating (prodromal) health concerns⁹.

The differentiation of true risk factors or triggers of MS from prodromal features will require further investigation. This process is also complicated by the fact that the length of the MS prodromal phase is unknown but a pragmatic approach can be taken to minimize the risk of reverse causation when studying risk factors. For example, the date of MS symptom onset or of the first clinical demyelinating event can be used to demarcate the outcome rather than the date of diagnosis and a clear gap can be left between this outcome and the period during which a person is considered at risk^74–77. These approaches can increase the confidence that a true risk factor is being examined rather than a prodromal feature.

In addition, some aspects are highly likely to be a feature of the MS prodrome as well as a risk factor or trigger of MS. For example, low sunlight exposure and/or low serum levels of vitamin D could be a true risk factor for MS in some populations but could also be a feature of the MS prodrome. As individuals enter the prodromal phase, the deterioration of health and increased physician and hospital visits could reduce time spent outside, thus reducing sunlight exposure and lowering serum levels of vitamin D⁷⁸. More work is therefore needed to determine which factors might be risk factors and which are prodromal features and to understand when and in whom these factors are most clinically relevant.

Refining identification of the prodrome.

The process of identifying the clinical features of the MS prodrome is likely to be iterative — identification of possible clinical features of the prodrome will inform the better characterization of the early manifestations of clinical MS. Currently, we could be classifying true early manifestations of clinical MS as prodromal features, in which case we could be missing opportunities for prompt recognition of clinical demyelinating syndromes and MS. This could be particularly true when individuals are assessed in primary care or by non-specialists who are less familiar with the clinical features of MS. Conversely, care is needed to avoid over diagnoses and over treatment of atypical clinical syndromes^65,79, which could cause unnecessary harm. Given that the putative prodromal features of MS identified thus far are relatively common and often non-specific, the identification of MRI and/or laboratory biomarkers is needed to increase specificity for prodromal MS and develop a reliable definition of the prodromal phase.

Earlier recognition and management of MS.

Knowledge of the MS prodrome could facilitate the earlier recognition and management of the disease. The development of research criteria for prodromal MS that include a combination of clinical features and biomarkers with high sensitivity and specificity could enable reliable identification of individuals in the prodromal phase. This ability has clear implications for individuals who could benefit from enrolment in trials of interventions designed to prevent clinical MS and/or offer neuroprotection to mitigate future disability (BOX 1).