The Prevalence of Pediatric Myalgic Encephalomyelitis/Chronic Fatigue Syndrome in a Community‑Based Sample

Abstract

Background

Most pediatric prevalence studies of myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) have been based upon data from tertiary care centers, a process known for systematic biases such as excluding youth of lower socioeconomic status and those less likely to have access to health care. In addition, most pediatric ME/CFS epidemiologic studies have not included a thorough medical and psychiatric examination. The purpose of this study was to determine the prevalence of pediatric ME/CFS from an ethnically and sociodemographically diverse community-based random sample.

Method

A sample of 10,119 youth aged 5–17 from 5622 households in the Chicagoland area were screened. Following evaluations, a team of physicians made final diagnoses. Youth were given a diagnosis of ME/CFS if they met criteria for three selected case definitions. A probabilistic, multi-stage formula was used for final prevalence calculations.

Results

The prevalence of pediatric ME/CFS was 0.75%, with a higher percentage being African American and Latinx than Caucasian. Of the youth diagnosed with ME/CFS, less than 5% had been previously diagnosed with the illness.

Conclusions

Many youth with the illness have not been previously diagnosed with ME/CFS. These findings point to the need for better ways to identify and diagnose youth with this illness.

Introduction

Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) involving 6 or more months of fatigue and other symptoms (Fukuda et al. 1994; IOM 2015; Jason et al. 2006; Reeves et al. 2003) significantly impairs various aspects of children’s and adolescents’ lives, including physical functioning, school attendance and performance, and extracurricular activities (Josev et al. 2017; Torres-Harding et al. 2006; Walford et al. 1993). Krilov et al. (1998) found that only 14% of adolescents with ME/CFS attended school regularly, and Dowsett and Colby (1997) found ME/CFS to be the most common cause of prolonged medical leave from school among adolescents, and many continue to have symptoms and do not return to their premorbid level of functioning (Brown et al. 2012).

Among the basic issues under dispute has been the prevalence of pediatric ME/CFS in the population (Jordan et al. 1998). In the US, the Centers for Disease Control and Prevention (CDC) conducted the first pediatric ME/CFS surveillance study (Gunn et al. 1993) and estimated that, among adolescents aged 12–17, 2.7 per 100,000 had ME/CFS, indicating that ME/CFS was a relatively rare disorder among youth (Dobbins et al. 1997). A major limitation of the Gunn et al. study and a comparable study in Australia (Lloyd et al. 1990) was that the samples were obtained through physician referral; therefore, members of the community who do not or cannot access medical care were not included (and both samples of youth ultimately identified with ME/CFS were extremely small). The CDC conducted another study involving referrals from school nurses from junior and senior high schools in Wichita, Kansas, and Reno, Nevada, in which an ME/CFS prevalence of 24.0 per 100,000 was found for 12–17-year-olds (Dobbins et al. 1997). Again, the gatekeeper methodology, as well as reliance on previous diagnosis by a physician (rather than current evaluations), limited the validity of these findings. The CDC also conducted two community-based studies. The first was a study in San Francisco (Dobbins et al. 1997) that employed random digit dialing to households as a means of identifying children and adolescents with ME/CFS-like illness. In this study of adolescents aged 12–17, 116.4 per 100,000 were diagnosed with ME/CFS-like conditions. A second, similar CDC prevalence study was carried out by Jones et al. (2004). It too involved random digit dialing of the residents of Wichita, Kansas, and this group reported a prevalence of ME/CFS-like illness of 338 per 100,000. Neither of these studies, however, included a medical evaluation.

In the Netherlands, Nijhof et al. (2011) reported an ME/CFS prevalence rate of 111 per 100,000 in a survey of general practitioners, but this study only had a 41% participation rate among physicians, and again there was no medical evaluation. Jordan et al. (2000) found a ME/CFS prevalence rate of 60 per 100,000 or 0.06%, but in this study significant implementation issues occurred including a considerable amount of time that elapsed between the telephone screen and medical examination. In Norway, Bakken et al. (2014) identified cases of ME/CFS in the Norwegian Patient Register for the years 2008–2012. For youth aged 10–19, they estimated the incidence of ME/CFS at 43 per 100,000. However, this registry had cases that came from a large number of different hospitals and the criteria used to diagnose ME/CFS might have varied. A more recent study in Great Britain by Collin et al. (2016) found a prevalence of ME/CFS of 1900 per 100,000 among 16-year-olds, but data were based on self-reports without a medical examination. In Rimes et al.’s (2007) study of British general practitioners, the incidence rate of pediatric ME/CFS was estimated at 500 per 100,000. Therefore, the prevalence of pediatric ME/CFS is still unclear.

Based on past epidemiologic studies which may have had methodological flaws, as reviewed above, descriptions of pediatric ME/CFS such as those of Arav-Boger and Spirer (1995) described the usual patient as being previously athletic and ambitious, and upper middle-class. These findings resembled how adults with ME/CFS were described following the first generation of adult epidemiologic studies (Reyes et al. 2003). In an effort to control for deficiencies encountered in prior pediatric ME/CFS epidemiologic studies as outlined above, the current investigation collected data from a community-based, ethnically and sociodemographically diverse sample, and provided test positive and test negative control youth a medical and psychological examination. The authors expected these methods would generate a more accurate estimate of pediatric ME/CFS prevalence. It was also hypothesized that, as with the adult community-based ME/CFS literature (Jason et al. 1999; Reeves et al. 2007), the majority of identified pediatric cases would not have been previously diagnosed with ME/CFS.

Method

Institutional Review Board approval was obtained before launching the study, and informed consent was appropriately obtained. There are no potential conflicts of interest present in the form of grants, employment by, consultancy for, shared ownership in, or any close relationship with, an organization whose interests, financial or otherwise, that could be affected by the publication of this article.

Interviewer Selection and Training

Telephone interviewers were volunteers who underwent an interview process and training on how to navigate the online screening system, talk with participants, and gather information effectively. Interviewers read a script and were constantly supervised. Calls were made from 9 am to 9 pm every day of the week. Interviewers were instructed to stop calling a phone number after seven attempts with no answers and a scripted voicemail was left during the first and fourth call.

Procedures

The first stage of the study involved calling households in the greater Chicago area. Phone numbers and addresses were obtained from InfoUSA. Postcards were sent to the home addresses provided, with information about the study and contact information if those who received the postcards were interested in participating. A parent/caretaker was then contacted over the phone from each household in order to confirm that there was a youth aged 5–17 in the household and to gain permission to screen for pediatric ME/CFS-like illness.

The parent/caretaker respondent was then interviewed regarding the presence or absence of fatigue in each child and adolescent in the household; all eligible children and adolescents within each household were screened. The Pediatric ME/CFS Screening Questionnaire was administered to respondents to screen for ME/CFS-like profiles among children and adolescents (Jason and Sunnquist 2018). This screening questionnaire consists of three parts. First, there are questions designed to determine if any of the children or teenagers in the household were experiencing significant fatigue. The second part assessed whether any of the children experienced a disruption in their school activities or performance due to fatigue or cognitive difficulties. These initial questions were broad because we wished to cast a wide net to increase sensitivity. The third part of the questionnaire presented a list of ME/CFS-related symptoms in children proposed by Bell (1995), Jason et al. (2006), and others (Fukuda et al. 1994; IOM 2015; Carruthers et al. 2003; Jason et al. 2010). The questionnaire also included questions about the youth’s medical history, current symptoms, and symptom onset.

Children and adolescents with no exclusionary medical conditions, who screened positive for either significant fatigue or school/learning/memory problems, had substantial reductions in functioning, and three or more ME/CFS Fukuda et al. (1994), IOM (2015), or Carruthers et al. (2003) symptoms were considered screen-positive and selected for full evaluation in Stages 2 and 3. Additionally, screen-negative control participants were brought in for Stages 2 and 3. These control individuals were selected based on interest in participating in the study and demographic matching for gender, age, and ethnicity with screen positive participants.

For stage two of the study, those caretakers/parent(s) and youth who agreed to participate and met criteria visited Lurie Children’s Hospital for a structured psychiatric interview, psychosocial assessment, and medical evaluation. The parents were informed that this was a study of fatigue in children and adolescents and that youth with and without fatigue were being selected to participate. At the beginning of the visit, a consent form was completed by the parent/guardian. Signed or verbal assent was obtained for adolescents as appropriate based on age. Children and their parents were compensated for their participation.

Prior to the in-person session, the parent/caretaker and youth completed the pediatric version of the DePaul Symptom Questionnaire (the DSQ; Jason and Sunnquist 2018), the Child Health Questionnaire (CHQ; Landgraf et al. 1996), the Autonomic Symptom Checklist (Sletten et al. 2012), and the Fatigue Severity Scale (FSS; Krupp et al. 1989). See below for more details.

The Schedule for Affective Disorders and Schizophrenia for School Aged Children Present and Lifetime Version (K-SADS-PL; Kaufman et al. 1997; see below for more information) was also separately administered to both the child and parent/caretaker during the in-person appointment. Research assistants, supervised by an advanced doctoral trainee with a master’s degree, administered the interview to the child and parent. Next, the youth were provided a medical interview and a complete physical examination. The examination included comprehensive questioning regarding symptoms of ME/CFS [e.g., asking about all of the Fukuda criteria (see below)]. As part of the physical, every subject was examined for lymphadenopathy and trigger points, a Beighton score for joint hypermobility was obtained, and a passive standing test was performed. All lab tests necessary to rule out medical disorders that could explain fatigue, such as anemia or hypothyroidism, were obtained, as per Fukuda et al. (1994).

Measures

DePaul Symptom Questionnaire (DSQ)

The DSQ is a self-report measure of ME and CFS symptomatology and illness history (Jason and Sunnquist 2018). This questionnaire provides a standardized method for assessing the dimensions of various case definitions, including the Fukuda et al. (1994) CFS criteria, Canadian Clinical criteria (Carruthers et al. 2003), and Institute of Medicine (IOM 2015) criteria. We obtained case definition fulfillment based on the child and parent/caretaker reports using a version of the DSQ developed for pediatric use (Jason et al. 2006, 2009; Jason and Sunnquist 2018). Participants rated each symptom’s frequency over the past 3 months on a 5-point Likert scale: 0 = “none of the time,” 1 = “a little of the time,” 2 = “about half the time,” 3 = “most of the time,” and 4 = “all of the time.” Likewise, participants rated each symptom’s severity over the past 3 months on a similar 5-point Likert scale: 0 = “symptom not present,” 1 = “mild,” 2 = “moderate,” 3 = “severe,” 4 = “very severe.” Frequency and severity scores were multiplied by 25 to create 100-point scales. The 100-point frequency and severity scores for each symptom were then averaged to create one composite score per symptom. Subsequently, domain composite scores for these factors were created by taking the 100-point score for each symptom’s frequency and severity within the domain and averaging them for one 100-point domain composite score. Fatigue frequency and severity was measured in the same manner and respondents were also asked to specify how many months they had experienced fatigue at the time of reporting (this information was used to determine if fatigue lasted 6 or more months for the case definitions).

The DSQ has evidenced good test–retest reliability among both patient and control groups (Jason et al. 2015) and has a three-factor structure that evidences good internal consistency (Brown and Jason 2014). Gleason et al. (2018) used this instrument to validate cut-off scores for substantial reductions in functioning with a young adult sample. Strand et al. (2016) found a sensitivity of 98% when comparing the agreement between a physicians’ diagnosis of ME/CFS using the Canadian Consensus Criteria and the DSQ’s assessment of this case definition. Murdock et al. (2016), an independent group using the DSQ, found that it demonstrated excellent internal reliability and that among patient-reported symptom measures, it optimally differentiated between patients and controls. The DSQ is available in the shared library of Research Electronic Data Capture (REDCap; Harris et al. 2009).

K‑SADS‑PL (Schedule for Affective Disorders and Schizophrenia for School Aged Children Present and Lifetime Version; Kaufman et al. 1997)

This structured interview schedule for children and adolescents has good psychometric properties. Psychiatric evaluation is essential to rule out psychiatric diagnoses that would preclude a diagnosis of ME/CFS and to differentiate psychiatric disorders from ME/CFS (Pelcovitz et al. 1995). Participants were asked about past and current symptoms of anxiety disorders, among other psychiatric illnesses, and their responses to each symptom were rated on a 3-point scale.

Child Health Questionnaire

The parent and child or adolescent completed the Child Health Questionnaire (CHQ; Landgraf et al. 1996), an instrument that assesses physical and psychosocial well-being. The instrument has both an 87-item child form (CHQ-CF87) and a 50-item parent form (CHQ-PF50). The scales measured by both forms include physical functioning, role/social (emotional, behavioral, physical), bodily pain, general health perceptions, and self-esteem. The parent form has two additional scales, “Parent Impact-Emotional” and “Parent Impact-Time.” Internal consistency/reliability coefficients are good for the CHQ (0.84 for the child form; ranging from 0.80 to 0.88 for the parent form, depending upon the parents’ SES). Furthermore, the CHQ evidences acceptable validity. Discriminant validity of the subscales is excellent, as the scales can discriminate children with clinically defined conditions from a control sample (Landgraf et al. 1996).

Autonomic Symptom Checklist

The Autonomic Symptom Checklist (Sletten et al. 2012) is a 31 item measure that assesses extent of symptoms (e.g., “On average, how frequently do you feel dizzy or lightheaded?) and presence of symptoms [e.g., “Have you ever fainted (become unconscious and fell down)?]. The ASC was developed from the Autonomic Symptom Profile (Suarez et al. 1999), which has been validated for ME/CFS.

Fatigue Severity Scale (FSS)

Krupp et al.’s (1989) Fatigue Severity Scale includes nine items rated on seven-point scales and is sensitive to different aspects and gradations of fatigue severity. Most items on the Krupp fatigue scale are related to behavioral consequences of fatigue. This scale can discriminate between individuals with ME/CFS, MS, and primary depression (Pepper et al. 1993). In addition, the FSS has been normed on a sample of individuals with MS, Lupus, and healthy controls. Higher scores indicate more fatigue.

Participants

Of the 147,954 phone numbers, 93,989 households never answered their phones and did not return multiple voicemails, and 10,920 phone numbers were inactive. Of the 43,045 phone numbers reached, 20,011 were ineligible due to not having children between the ages of 5–17 in the household or because it was a business number, 12,372 were not interested in the study, and 5040 requested that we call back at a later time, but did not answer our subsequent phone calls. We were able to screen the remaining 5622 households, and survey data were gathered for 10,119 children and adolescents. Of these 10,119 children, 865 were potentially screened positive with either significant fatigue or school/learning/memory problems, and three or more ME/CFS Fukuda et al. (1994), IOM (2015), or Carruthers et al. (2003) symptoms. Of all these potentially screened positive youth, 298 were later classified as screen positive as they reported no exclusionary medical conditions and had substantial reductions in functioning. Of those who screened positive, 165 (55.4%) were able to complete Stage 2; others were not able to make the trip to the hospital or their schedule did not allow them time to do this stage of the study. Of those who screened negative, 243 were eligible to participate in Stage 2 based on demographic matching protocols, and 42 (17.3%) were recruited for Stage 2 as control participants. Of the children who participated in Stage 2, two children had current diagnoses of ME/CFS and both were in the test positive group (i.e., their prior ME/CFS diagnoses were confirmed by the physician review panel of the present study).

Diagnosing ME/CFS

At the end of Stage 3, a team of physicians were responsible for making final diagnoses. Two physicians independently rated each youth according to the ME/CFS case definitions described below.

Case Definitions

1. Fukuda et al. (1994) criteria: To be diagnosed using the Fukuda et al. (1994) criteria, participants needed to experience persistent or relapsing fatigue for a period of six or more months concurrent with at least four of eight somatic symptoms that did not predate the fatigue. Substantial reductions in functioning were measured by the Child Health Questionnaire (CHQ) (Landgraf et al. 1996) subscales, which assessed the child’s level of functioning in a variety of domains based on both parent-report and child self-report (e.g., “During the past 4 weeks, has it been difficult for you to get around your school, neighborhood, or playground due to health problems? 1 = “No, not difficult”, 4 = “Yes, very difficult”). In addition, to meet criteria, the youth could not have any exclusionary medical or psychiatric illnesses, as defined by Reeves et al. (2003).

2. Pediatric criteria (Jason et al. 2006): This case definition is modeled after the Canadian Clinical Criteria case definition (Carruthers et al. 2003). To be diagnosed with ME/CFS via the Pediatric criteria (Jason et al. 2006), participants needed to have unexplained, persistent, or relapsing chronic fatigue not the result of ongoing exertion and not substantially alleviated by rest. The CHQ was used to determine substantial reductions in previous levels of educational, social, and personal activities. This case definition has been used by other investigators; for example, Josev et al. (2017) used these criteria to select 21 adolescents with ME/CFS, who were found to have significantly longer objective sleep onset latency, time in bed, total sleep time, and significantly poorer subjective sleep quality compared with healthy adolescents. Using this case definition, Jason et al. (2009) were able to differentiate those with more severe versus moderate symptoms, and Jason et al. (2010) found fewer ME/CFS patients would be misdiagnosed using the Pediatric ME/CFS case definition than those of Fukuda et al.(1994). For the purposes of this study, a 6-month threshold for fatigue was used in order to meet this criterion to maintain continuity of fatigue duration across the three case definitions used for evaluation.

3. IOM (2015) criteria: The IOM (2015) was operationalized by having youth meet the following four criteria: (1) Substantial reduction in activity as assessed by the CHQ; (2) Post-exertional malaise as assessed using the following criteria: soreness after mild activity, drained/sick after mild activity, minimum exercise makes tired, muscle weakness, dead/heavy feeling after exercise, and mentally tired after the slightest effort; (3) Sleep dysfunction as assessed using the following: unrefreshing sleep, problems staying asleep, problems falling asleep, waking up early, and need to nap daily; and finally, (4) Neurocognitive impairment or orthostatic intolerance (neurocognitive symptoms included difficulty paying attention, difficulty expressing thoughts; orthostatic intolerance items include feeling unsteady on your feet, like you might fall; dizziness).

Diagnosis of ME/CFS

Two reviewing physicians had access to all information gathered on each participant during each of the three phases of the study including results from the physical exam. In the clinical impression following the stage 3 physical exam, the examining physician (BZK) did not make a final diagnostic conclusion regarding whether a participant has ME/CFS, because he did not have access to all information (e.g., psychiatric diagnosis). Thus, information provided to the reviewers included physical exam results, the presence of exclusionary medical conditions, laboratory results, a summary of all parent and youth self-report measures, and the general clinical impression. Individuals evaluated as meeting ME/CFS criteria were given a final diagnosis of ME/CFS using the Fukuda et al. (1994) criteria with revisions recommended by Reeves et al. (2003), the pediatric ME/CFS criteria developed by an international study group (Jason et al. 2006), and the IOM (2015) clinical criteria.

Data Analysis

The prevalence of ME/CFS was calculated using the formula cited in Jason et al. (2012). This formula takes into account the actual number of participants who had ME/CFS as well as those who had a chance of diagnosis based on screening positive in Stage 1 but may not have participated in Stage 2, yielding a more encompassing prevalence estimate. The total number of respondents screened in Stage 1 is represented by N. The proportion of screened positives over the total number of screens in Stage 1 is represented by PI, and the proportion of screened negatives over the total number of screens in Stage 1 is represented by 1 − PI. The proportion of screened positives who were evaluated in Stage 2 and diagnosed with ME/CFS is represented by L1 and the proportion of screened negatives who were evaluated in Stage 2 and diagnosed with ME/CFS is represented by L2. This information was then used in the following formula to obtain the Prevalence P: P = L1 * PI + L2 * (1 − PI).

Chi square analyses were used to examine group differences between screen positive participants and screen positive non-participants first to determine whether there were any significant differences in gender, age, and race/ethnicity and whether equal prevalence could be assumed. Second, descriptive statistics, Chi square analyses, and t tests were used to examine differences in prevalence rates among groups and symptom endorsement between those diagnosed with ME/CFS and screen negative control participants.

Results

Table 1 presents frequency data for screen positive and screen negative participants as well as final diagnoses for ME/CFS. There were no significant differences between the screen positive subjects and screen negative controls in terms of gender, race/ethnicity, and age, as expected, as test negative control participants were invited to participate based on a demographic-matching process.

Table 1.

Data on participant selection and completion of the study

	Number of participants
	Screened positive	Screened negative
Completed phase one screen	865	9254
Selected for phase two of evaluation	298	243
Completed phase two of evaluation (physician review)	165	42
Final diagnosis of ME/CFS	42	0

Prevalence rates, using the formula delineated above, classified ME/CFS if youth met the Fukuda et al. (1994), IOM (2015), and Pediatric (Jason et al. 2006) case definitions.

$$\text{P }=\left(42/165\right)\ast \left(298/10,119\right)+\left(0/42\right)\ast \left(1-298/10,119\right)$$

$$\text{P }=(0.2545\ast 0.0294)+0$$

$$\text{P }=0.0075=750/100,000$$

Thus, the prevalence for ME/CFS in this community-based pediatric population was found to be 0.75% (95% confidence interval, 0.54–0.96%), or 750 per 100,000.

Table 2 presents prevalence estimates of ME/CFS among varying sociodemographic groupings including gender, racial/ethnic identification, and age. In terms of gender, females were found to have a higher prevalence rate of ME/CFS than males. The prevalence rates for ME/CFS were also higher for Latinx and African American individuals when compared to Caucasian youth. Those aged 14–17 had higher prevalence rates than younger children.

Table 2.

Prevalence rates of ME/CFS (per 100,000 persons)

	Number of respondents	Individuals with ME/CFS	ME/CFS prevalence rate
Total	10,119	42	750 ± 109
	N	N	p ± SE
Gender
Female	4896	25	935 ± 174
Male	5191	17	582 ± 134
Race/ethnicitya
Caucasian	6835	23	632 ± 122
African American	1106	7	1108 ± 402
Latinx	1292	11	1277 ± 326
Age
5–8	1824	1	140 ± 135
9–11	2223	6	453 ± 176
12–14	2927	16	816 ± 195
15–17	3109	19	1249 ± 264

^aOther ethnicities listed included: 320 Multiracial (Not Hispanic/Latino); 24 American Indian/Alaskan Natives; 5 Native Hawaiian/Pacific Islander; and 307 preferred not to respond to this question (there were no cases of ME/CFS in these groups). One Asian American met ME/CFS criteria, but given the small sample size of Asian Americans in the sample (n = 230), calculating prevalence was not appropriate for this group

Table 3 includes sociodemographic data of those with ME/CFS and controls. No significant differences were found between participants with ME/CFS and controls in regards to gender, race/ethnicity, and age, again as expected based on our study design and matching process.

Table 3.

Demographic characteristics of stage two participants

	ME/CFS (n = 42) M (SD)	Control (n = 42) M (SD)
Age	14.0 (2.5)	13.5 (2.2)
	% (n)	% (n)
Gender
Female	59.5 (25)	59.5 (25)
Male	40.5 (17)	40.5 (17)
Race
African American	16.7 (7)	19.0 (8)
White/Caucasian	54.8 (23)	45.2 (19)
Asian American	2.4 (1)	7.1 (3)
Latinx	26.2 (11)	19.0 (8)
Multiracial (not of Latinx origin)	0 (0)	9.5 (4)
Family income
$150,000 or more	31.0 (13)	40.4 (17)
$100,000-$149,999	19.0 (8)	21.4 (9)
$50,000-$99,999	21.4 (9)	19.0 (8)
$25,000-$49,999	16.7 (7)	11.9 (5)
< $24,999	7.1 (3)	2.4 (1)
Prefer not to respond	4.8 (2)	4.8 (2)
Highest degree or level of education
Graduate or professional degree	35.7 (15)	42.9 (18)
Standard college degree	35.7 (15)	42.9 (18)
Partial college or specialized training	19.8 (8)	7.1 (3)
High school or GED	7.1 (3)	4.8 (2)
Some high school	0 (0)	2.4 (1)
Less than high school	2.4 (1)	0 (0)
Employment status
Employed full-time	42.9 (18)	64.3 (27)
Employed part-time	19.0 (8)	16.7 (7)
Homemaker	31.0 (13)	14.3 (6)
Unemployed/looking for work	7.1 (3)	2.4 (1)
Student	0 (0)	2.4 (1)

Table 4 presents the mean number of symptoms reported, duration of fatigue, and findings from the FSS and the Autonomic Symptom Checklist for those with ME/CFS versus screen negative controls. Symptoms are presented using the threshold frequency of at least half of the time and threshold severity of moderate or greater. Significant differences were found between participants with ME/CFS and controls in regards to number of symptoms reported, duration of fatigue, fatigue scores on the FSS, Autonomic Symptom Checklist data, and almost all symptoms, as expected.

Table 4.

Child-report symptoms among ME/CFS and screen negative controls

	ME/CFS (n = 42) M (SD)	Control (n = 42) M (SD)	Sig.
Mean number of symptoms	17.3 (9.0)	2.5 (4.7)	***
Months of fatigue	27.9 (35.1)	1.3 (.6)	***
Autonomic symptoms checklist total score	31.0 (18.5)	10.2 (9.2)	***
Fatigue severity scale total score	43.1 (12.1)	19.5 (13.5)	***
	%	%	Sig.
Fatigue	90.5	9.5	***
Post Exertional Malaise (at least one symptom)	83.3	7.1	***
Drained/Sick after mild activity	40.5	4.8	***
Minimum exercise makes tired	61.9	4.8	***
Soreness after mild activity	38.1	0	***
Dead/Heavy feeling after exercise	45.2	4.8	***
Mentally tired after slightest effort	52.4	4.8	***
Sleep (at least one symptom)	95.2	23.8	***
Unrefreshing sleep	90.5	19.0	***
Problems staying asleep	33.3	2.4	***
Waking up early	28.6	4.8	**
Problems falling asleep	57.1	4.8	***
Need to nap daily	45.2	2.4	***
Pain (at least one symptom)	90.5	19.0	***
Muscle pain	50.0	9.5	***
Eye pain	23.8	2.4	**
Joint pain	40.5	7.1	**
Chest pain	26.2	0	***
Stomach pain	38.1	7.1	**
Headaches	64.3	7.1	***
Muscle twitches	14.3	2.4	*
Nausea	31.0	0	***
Upset stomach	45.2	7.1	***
Ringing in the ears	21.4	7.1
Vomiting	2.4	0
Neurocognitive (at least one symptom)	85.7	26.2	***
Difficulty understanding	42.9	7.1	***
Absent-mindedness	52.4	2.4	***
Slowness of thought	33.3	2.4	***
Problems remembering things	52.4	4.8	***
Difficulty paying attention	61.9	14.3	***
Difficulty finding the right word	38.1	7.1	**
Can only focus on one thing at a time	54.8	4.8	***
Loses train of thought	42.9	9.5	**
Trouble with math or numbers	50.0	16.7	**
Autonomic (at least one symptom)	61.9	11.9	*
Unsteady on feet	19.0	2.4	*
Shortness of breath	38.1	4.8	***
Dizziness	38.1	4.8	***
Loss or gain of weight	19.0	9.5
Irregular heartbeat	14.3	0	**
Neuroendocrine (at least one symptom)	61.9	19.0	***
Feeling hot/cold for no reason	40.5	2.4	***
High temperature	14.3	0	*
Low temperature	7.1	0
Chills/Shivers	16.7	0	**
Sweaty hands	11.9	9.5
Night sweats	19.0	4.8	*
Loss of appetite	16.7	7.1
Immune (at least one symptom)	50	19.0	**
Fever	11.9	2.4	*
Sore throat	14.3	2.4
Tender/Sore lymph nodes	7.1	2.4
Sensitivity to smells/foods/chemicals	26.2	0	***
Rashes	11.9	0	*
Allergies	26.2	14.3

Symptoms reported at least half of the time and of moderate or greater severity

^*p < 0.05;

^**p < 0.01;

^***p < 0.001

Discussion

This study assessed the prevalence of pediatric ME/CFS among youth from a sample unbiased by chronic illness or help-seeking behaviors. In this ethnically and sociodemographically diverse community-based sample, the ME/CFS prevalence rate was estimated to be 750 per 100,000 children. Previous epidemiological pediatric ME/CFS outcomes have found rates ranging from 2.7 to 1900 per 100,000 children (Gunn et al. 1993; Chalder et al. 2003; Farmer et al. 2004; Jones et al. 2004; Jordan et al. 2006; Rimes et al. 2007; Nijhof et al. 2011; Bakken et al. 2014; Collin et al. 2016). However, many of these studies relied on physician referrals, which excludes those who cannot or do not access medical care, and this exclusion may have reduced previous prevalence estimates. In addition, not all previous studies included a physical examination and screening laboratory evaluation to exclude other diagnoses, which may have inflated previous estimates. Of the 42 youth diagnosed with ME/CFS, only 2 or 4.8% had been previously diagnosed with ME/CFS; thus, most youth with this illness are not diagnosed as such in the community.

Substantially higher prevalence rates of ME/CFS were found for females, individuals from racial/ethnic minorities, and older children. In past research, females have also been found to have higher rates of ME/CFS than males (Lloyd et al. 1990; Jordan et al. 2006; Bakken et al. 2014). While previous pediatric prevalence studies have generally not reported prevalence data for racial/ethnic groups, adult community-based epidemiological studies have found racial/ethnic minorities to have higher prevalence rates of ME/CFS (Jason et al. 1999; Reeves et al. 2007). Higher rates among minorities could be due to them having less access to adequate health care. Higher rates among older youth compared to younger youth could be due to endocrinological changes and psychological issues inherent in adolescence; younger youth might also be exposed to fewer environmental and bio-logical precipitating causal factors. Jason et al. (2000) found that those patients with ME/CFS who were older had higher frequencies of symptoms and were more severely disabled. Conceivably, those who are older might be less physically fit or have had more opportunities to experience other physical illnesses or stressors that can cause fatigue. It is important for future research to explore reasons for increased ME/CFS risk among these groups.

For the youth identified with ME/CFS, 50% had family incomes of $100,000 or greater, and 71% of parents had a college or higher education. Current census data for the greater Chicago area estimates that 31.3% of the population has a family income of $100,000 or greater (U.S. Census Bureau, 2017a) and 37.4% of the population has a college degree or higher education (U.S. Census Bureau 2017b). Discrepancies between the demographic characteristics of this sample and the greater population could suggest that families of higher education or income may have been more willing to answer the phone or participate in the study. Additionally, these families may have been more likely to be able to take the time to attend Stage 2 as opposed to those of lower education or income status who may not have been able to take off of work or travel to the hospital.

Finally, less than 5% of the identified youth with ME/CFS had been previously diagnosed with ME/CFS; this finding is comparable to adult community-based findings regarding the majority of adults also being not identified previously (Jason et al. 1999). These findings point to the need for better ways to identify youth with this illness and to develop appropriate rehabilitation interventions for them.

Our study collected child and parent-report data on measures including the DSQ and CHQ, and this information was included as part of the physician review process. Our symptom data in Table 4 is based on youth data, but comparable differences were also found when inspecting data from parents. Those with ME/CFS had more symptoms overall than those from the control group, as expected. Those with ME/CFS had an average of over 17 symptoms, lasting over 2 years, and with over 85% having classic ME/CFS symptoms such as severe and frequent fatigue, post-exertional malaise, sleep, pain and neurocognitive symptoms. Discrepancies in symptom reporting among participants may influence prevalence rates, as symptoms may be over- or underreported depending on the respondent (Kennedy et al. 2010; Waters et al. 2003). This is something to consider when understanding how a diagnosis of ME/CFS is made by reviewing physicians as parent and child reports are integrated into the diagnostic picture.

As indicated in Table 4, 90.5% of youth with ME/CFS met the frequency and severity criteria for fatigue based on child-report; however, 4 youth with ME/CFS did not meet these criteria. Given the prominence of the symptom of fatigue for a diagnosis of ME/CFS, it is useful to more closely review these cases. The physician panel evaluated all 4 youth as having ME/CFS as the parent report on the DSQ did indicate fatigue at the threshold level for all 4 cases and all children did endorse the presence of fatigue. Additionally, although they did not endorse threshold level fatigue on self-report measures, children are less likely than adults to report fatigue, as they may have difficulty verbalizing the experience (Bell 1995; Lapp 2006). The physicians and research staff asked detailed questions related to fatigue during participants’ medical and psychiatric interview, and participants’ responses to this information is included in summary reports for the physician panel. These children also complained of post-exertional malaise, memory/concentration, and sleep disturbance.

Both adult and pediatric ME/CFS epidemiological studies have found a wide range of prevalence rates, although in general, pediatric rates have been lower. There have also been some studies that suggested that children’s recovery rates from ME/CFS may be quicker and higher than adults, which could explain why pediatric prevalence rates have tended to be lower (Patel et al. 2003; Devendorf et al. 2019; Rowe 2019). Adults have had more time to be exposed to more illnesses, stressors and medications, which also could be responsible for higher prevalence rates.

The lack of an agreed upon case definition has complicated efforts to estimate prevalence rates of pediatric ME/CFS. Three widely used case definitions in the field were utilized in this study, with diagnostic criteria ranging from a more broad perspective (Fukuda et al. 1994) to more stringent criteria (Jason et al. 2006; IOM 2015). For the current study, an individual was diagnosed with ME/CFS only if they met all three case definitions. We thus used a conservative strategy to accurately identify cases. The pediatric criteria (Jason et al. 2006) modeled after the Canadian Consensus Criteria (Carruthers et al. 2003) were found to be the most stringent.

Our study has several limitations. Although this was a random community-based sample, we were only able to reach a small percentage of those we attempted to contact. Using telephones is becoming an increasingly difficult method to recruit subjects. Our sample may have been biased in that it involved only those who answered the phone and agreed to be interviewed and seemed to involve a greater percentage of families with higher educational attainment. There is also an assumption that the proportion of those evaluated and found to have ME/CFS is the same as the proportion who have ME/CFS in the population screened. We did find that there was differential attendance between those that screened positive (55% completed stage 2) and those that screened negative (under 18% completed stage 2). Although this may suggest that those who had ME/CFS were more likely to complete the study, we only attempted to recruit a matched sample of 42 controls. As indicated in Table 2, the prevalence rates do have a wide standard error. Finally, the 6 month requirement for a diagnosis is not required in every country; in some cases only 3 months is required (Rowe et al. 2017). Therefore, our prevalence rate may not be generalizable to those sites where only 3 months of fatigue is required, and therefore the prevalence in using the 3 month criterion may be higher.

In summary, we determined the prevalence rate of pediatric ME/CFS in the community to be 0.75%. In addition, it appears that females, those from Latinx and African American ethnicities have higher rates than Caucasians. Of particular importance, over 95% of identified youth had not previously been diagnosed. Autonomic, fatigue, and symptom data suggest that those who have this illness have significant and multiple impairments, and thus represent a high-risk group that needs to be both diagnosed and appropriately treated.