Source: Journal of DisabilityPolicy Studies Online

Date:   October 21, 2008

URL:    http://jdps.sagepub.com

        http://dps.sagepub.com/cgi/content/abstract/1044207308325995v1

Evaluating the Centers forDisease Control's Empirical Chronic Fatigue

Leonard A. Jason(*), NatashaNajar, Nicole Porter, Christy Reh,

DePaul University, Chicago,Illinois

* Address correspondence toLeonard A. Jason, PhD, Director, Center for

  Community Research, DePaul University, 990 W. Fullerton Avenue,Suite 3100,

  Chicago, IL 60614; e-mail: Ljason@depaul.edu.

Abstract

The Centers for Disease Controland Prevention (CDC) recently developed an empirical case definition thatspecifies criteria and instruments to diagnose chronic fatigue syndrome (CFS)in order to bring more methodological rigor to the current CFS case definition.The present study investigated this new definition with 27 participants with adiagnosis of CFS and 37 participants with a diagnosis of a Major DepressiveDisorder. Participants completed questionnaires measuring disability, fatigue,and symptoms. Findings indicated that 38% of those with a diagnosis of a MajorDepressive Disorder were misclassified as having CFS using the new CDCdefinition. Given the CDC's stature and respect in the scientific world, thisnew definition might be widely used by investigators and clinicians. This mightresult in the erroneous inclusion of people with primary psychiatric conditionsin CFS samples, with detrimental consequences for the interpretation of epidemiologic,etiologic, and treatment efficacy findings for people with CFS.

Keywords: chronic fatiguesyndrome; empirical case definition; Centers for Disease Control andPrevention; Fukuda criteria; Major Depressive Disorders

Chronic fatigue syndrome (CFS) isa disabling chronic illness that has been defined by a consensus-based approachby Fukuda et al. (1994). This case definition specifies that individuals withthis illness must have 6 or more months of chronic fatigue of new or definiteonset, which is not substantially alleviated by rest, is not the result ofongoing exertion, and  results insubstantial reductions in occupational, social, and personal activities. Inaddition, to be diagnosed with this illness, individuals must have four or moresymptoms (i.e., sore throat, lymph node pain, muscle pain, joint pain, post exertionalmalaise, headaches of a new or different type, memory and concentrationdifficulties, and un-refreshing sleep) that persist 6or more months sinceonset. Although the Fukuda et al. case definition continues to be widely used,several articles have identified difficulties that this case definitioncontinues to pose to clinicians and researchers (Jason, King, et al., 1999;Reeves et al., 2003). For example, the Fukuda et al. case definition did notspecify which instruments to use and did not provide empirically derived cutoffpoints and scoring guidelines to diagnose CFS.

The Centers for Disease Controland Prevention (CDC) has now developed an empirical case definition for CFSthat involves assessment of symptoms, disability, and fatigue (Reeves et al.,2005). The new CDC empirical case definition assesses disability using theMedical Outcomes Survey Short-Form-36 (Ware, Snow, & Kosinski, 2000),assesses symptoms using the Symptom Inventory (Wagner et al., 2005), andassesses fatigue using the Multidimensional Fatigue Inventory (Smets, Garssen,Bonke, & DeHaes, 1995). The authors of this empirical case definition feelthat the specification of instruments and cutoff points would result in a morereliable and valid approach for the assessment of CFS. Using these newcriteria, the estimated rate of CFS has increased to 2.54% (Reeves et al.,2007), a rate that is about 10 times higher than prior CDC estimates (Reyes etal., 2003) and prevalence estimates of other investigators (Jason, Richman, etal., 1999).

It is of interest that the newCFS rates are within the range of several mood disorders. Mood disorders arethe most prevalent psychiatric disorders after anxiety disorders: For a majordepressive episode, the 1-month prevalence is 2.2%, and lifetime prevalence is5.8% (Regier, Boyd, & Burke, 1988). It  is at least possible that the increases in the United Statesare due to a broadening of the case definition and possible inclusion of caseswith primary psychiatric conditions. CFS and depression are two distinct disorders,however, even if they share a number of common symptoms. Including patientswith a primary psychiatric illness in the current CFS case definition couldconfound the interpretation of epidemiologic and treatment studies. MajorDepressive Disorder (MDD) is an example of a primary psychiatric disorder thathas some overlapping symptoms with CFS. Fatigue, sleep disturbances, and poorconcentration occur in both depression and CFS. It is important todifferentiate those with a principal diagnosis of MDD from those with CFS only.This is particularly important because it is possible that some patients withMDD also have chronic fatigue and four minor symptoms that can occur withdepression (e.g., unrefreshing sleep, joint pain, muscle pain, and impairmentin concentration). Fatigue and these four minor symptoms are also definingcriteria for CFS. It is possible that using this broadened new CFS empiricalcase definition (Reeves et al., 2005), some patients with a primary affectivedisorder could be misdiagnosed as having CFS. Some CFS investigators would notsee this as a problem because they believe that CFS is mainly a psychiatricdisorder and that distinctions between the two phenomena are superficial andmerely a matter of nomenclature. However, several CFS symptoms, includingprolonged fatigue after physical exertion, night sweats, sore throats, andswollen lymph nodes, are not commonly found in depression. In addition,although fatigue is t he principal feature of CFS, fatigue does not assumeequal prominence in depression (Friedberg & Jason, 1998; Komaroff et al.,1996). Moreover, illness onset with CFS is often sudden, occurring over a fewhours or days, whereas primary depression generally shows a more gradual onset.Individuals with CFS can also be differentiated from those with depression by  recordings of skin temperature levelsand electrodermal activity (Pazderka-Robinson, Morrison, & Flor-Henry,2004). Hawk, Jason, and Torres-Harding (2006) used discriminant functionanalyses to identify variables that successfully differentiated patients withCFS, MDD, and controls. Using percentage of time fatigue was reported,postexertional malaise severity, unrefreshing sleep severity,confusion/disorientation severity, shortness of breath  severity, and self-reproach to predictgroup membership, 100% were classified correctly. In summary, CFS anddepression are two distinct disorders, although they share a number of commonsymptoms. It is possible to appropriately differentiate MDD from CFS if oneuses appropriate measures.

It is still unclear whether thenew empirical case definition of CFS (Reeves et al., 2005) has inappropriatelyincluded cases of purely affective disorders, such as MDD. This study evaluatedwhether the CDC empirical case definition distinguished between persons withMDD and persons with CFS. By assessing samples with MDD and CFS, we hoped toclarify whether the CDC empirical case definition has been able to successfullydifferentiate those with MDD from those with CFS.

Method

Participants

We recruited a total of 64individuals, 27 with CFS and 37 with MDD. We obtained our sample ofparticipants with CFS from two sources: local CFS support groups in Chicago anda previous research study conducted at DePaul University. To be included in thestudy, participants were required to have been diagnosed with CFS, using theFukuda et al. (1994) diagnostic  criteria,by a certified physician and were required to currently meet CFS criteria usingthe Fukuda et al. criteria. We excluded individuals who had other currentpsychiatric conditions in addition to major depression or who reported havinguntreated medical illnesses (e.g., diabetes, anemia).

We solicited 37 participants witha diagnosis of MDD to participate in this study. We found participants fromthree sources: local chapters of the Depression and Bipolar Support Alliancegroup in Chicago; Craigslist, a free local classified ads forum that is communitymoderated; and online depression support groups. To be included in the study,all participants were required to have been diagnosed with MDD by a licensedpsychologist or psychiatrist. We excluded individuals who had other currentpsychiatric conditions in addition to MDD (e.g., bipolar, schizophrenia) andthose who reported having untreated medical illnesses.

Participants who met criteriacompleted questionnaires that are described below. Participants reported anyprevious physical and mental illnesses and the date of diagnosis as well ascurrent medications being taken to ensure that no other illness could accountfor the fatigue. We carefully screened participants to ensure that participantsfrom the MDD group did not have CFS as defined by the Fukuda et al. (1994)criteria.

Measures

Demographic variables.

We collected basic demographicvariables that included age, ethnicity, marital status, occupation, gender,work status, and educational level. The Medical Outcomes Survey Short-Form-36. This36-item instrument is composed of multi-item scales that assess functionalimpairment in eight areas: limits in physical activities (Physical Function),limits in one's usual role activities due to physical health (Role Physical),limits in one's usual role activities due to emotional health (Role Emotional),Bodily Pain, general health perceptions (General Health), vitality (Energy andFatigue), Social Function, and General Mental Health (Ware et al., 2000).Scores in each area reflect ability to function, and higher values indicatebetter functioning. Reliability and validity studies have demonstrated highreliability and validity in a wide variety of patient populations for thisinstrument (Stewart, Greenfield, Hays, et al., 1989). Based on the CDCempirical case definition (Reeves et al., 2005), the Medical Outcomes SurveyShort-Form-36 was used to assess disability (Wagner et al.,2005). According toReeves et al. (2005), significant reductions in occupational, educational,social, or recreational activities were defined as scores lower than the 25thpercentile on Physical Function (less than or equal to 70), or Role Physicalfunction (less than or equal to 50), or Social Function (less than or equal to75), or Role Emotional function (less than or equal to 66.7). A person wouldmeet the disability criterion for the empirical CFS case definition by showingimpairment in only one or more of these four areas (Reeves et al., 2005).

The CDC Symptom Inventory.

The CDC Symptom Inventoryassesses information about the presence, frequency, and intensity of 19fatigue-related symptoms during the past 1 month (Wagner et al., 2005). All 8of the critical Fukuda et al. (1994) symptoms were included as well as 11 othersymptoms (e.g., diarrhea, fever, sleeping problems, and nausea). For each ofthe 8 Fukuda et al. symptoms, participants were asked to report the frequency(1 = a little of the time, 2 = some of the time, 3 = most of the time, 4 = allof the time) and severity (the ratings were transformed to the following scale:.08 = very mild, 1.6 = mild, 2.4 = moderate, 3.2 = severe, 4 = very severe; seeNote 1). The frequency and severity scores were multiplied for each of the 8critical Fukuda et al. symptoms and were then summed. Participants having 4 ormore symptoms and scoring greater than or equal to 25 would meet symptomcriteria on this instrument according to the CDC empirical case definition(Reeves et al., 2005).

The Multidimensional FatigueInventory.

This instrument is a 20-item self-reportinstrument consisting of five scales: General Fatigue, Physical Fatigue,Reduced Activity, Reduced Motivation, and Mental Fatigue (Smets et al., 1995).Each scale contains four items rated from 1 to 5, with the scale score of 1meaning yes, that is true and the scale score of 5 meaning no, that is nottrue. Reeves et al. (2005)used the Multidimensional Fatigue Inventory tomeasure severe fatigue, and to do this, they used only two of the fivesubscales: General Fatigue and Reduced Activity. Using the CDC empirical casedefinition standards, severe fatigue was defined as greater than or equal to 13on General Fatigue or less than or equal to 10 on Reduced Activity.

Results 

Classification by CDC Empirical CaseDefinition Criteria

When using the CDC empirical case definitionto classify people with CFS, all 27 participants in the CFS-recruited group metcriteria for CFS. However, 14 additional individuals from the MDD group alsomet the new CDC criteria for CFS. That is, 38% of those with a professionaldiagnosis of major depression were misclassified as having CFS using the CDCempirical case definition.

Sociodemographic Variables

Participants were separated into threegroups: Those 27 diagnosed with CFS prior to this study and who met the newempirical CDC case definition of CFS, those 14 from the group with MDD meetingthe new empirical CDC case definition of CFS criteria (MDD/CFS), and those 23from the group with MDD not meeting the new empirical CDC criteria for CFS(MDD). Sociodemographic data were compared across all three groups ofparticipants using Pearson's chi^2 and analysis of variance (ANOVA; see Table1). Findings indicated a significant age effect, F(2,63)=3.25, p<.05. Theaverage age for the CFS group was significantly older than the MDD/CFS group.Furthermore, there were also significant differences in regard to work statusbetween groups, chi^2(6,N=64)=13.92, p<.05. More individuals in the CFSgroup were on disability as compared to the MDD/CFS group, chi^2(1,N=41)=4.11,p<.05.

Illness Classification by Standardized ClinicallyEmpirical Criteria

Medical Outcomes Survey Short-Form-36.

According to the CDC empirical casedefinition, participants are required to demonstrate functional impairmentwithin one of the four areas: Physical Function, Role Physical, Role Emotional,and Social Function. One-way ANOVA was used to assess the effect of physicalimpairment within four subscales of the Medical Outcomes Survey Short-Form-36for the three groups (CFS, MDD, and MDD/CFS). As seen in Table 2, there weresignificant effects for three of the subscales, but not social functioning.Using Tukey's honestly significant difference (HSD) post hoc test, significantdifferences were found for Role Physical; participants with CFS hadsignificantly lower scores compared to both the MDD group (p<.001) and theMDD/CFS group (p<.001). In regard to physical functioning, the participantswith CFS had significantly worse Physical Function impairment scores incompari- son to participants with MDD (p>.001) and participants with MDD/CFS(p<.001). Finally, for role emotional functioning, the MDD/CFS group scoredsignificantly lower on the  sole Emotionalscale than both the CFS (p<.001) and the MDD groups (p<.001). 

Examining Table 3, it is apparent that allthree illness groups met criteria for at least one of the four subscales andthus would meet the disability criteria for the empirical case definition ofCFS. It is clear that significantly more participants from the MDD and MDD/CFSgroups met Role Emotional criteria than the CFS group. However, if RolePhysical or Physical Functioning criteria were used as the sole criterion fordisability, significantly more participants within the CFS group would meet thedisability criteria than those in the MDD/CFS and MDD groups.

Symptom Inventory analysis.

There was a significant effect of the totalCFS symptom scores, F(2,61)=34.184, p<.001. The MDD group had the lowestmean score, indicating that this group did not likely meet criteria for CFS.The CFS group mean  score wasdirectionally but not significantly higher than the MDD/CFS group score. Tukeypost hoc tests indicated that the CFS and MDD/CFS groups scored significantlyhigher than the MDD group (p<.001). Examining Table 3, both the CFS andMDD/CFS groups had higher percentages of participants meeting CFS symptomcriteria than those in the MDD group. The fact that 100% of participants in theCFS and MDD/CFS groups met criteria for this index suggests that manyindividuals without CFS will meet these cutoff criteria for symptom frequencyand severity.

The Multidimensional Fatigue Inventory.

 TheCDC empirical case definition used the Multidimensional Fatigue Inventory tomeasure fatigue. There was a significant effect for General Fatigue, F(2,61)=4.89,p<.05, but no significant effect was found for Reduced Activity. Post hocanalysis using the Tukey HSD test revealed significant differences for GeneralFatigue. The MDD group scored significantly lower on the General Fatigue scalethan both the CFS (p<.01) and MDD/CFS groups (p<.01). Inspecting Table 3,all participants within the CFS and MDD/CFS groups met one of the fatiguecriteria. In addition, 87% of those in the MDD group also met one of thefatigue criteria. This again suggests that for the domain of fatigue, theempirical case criteria will select many individuals without CFS who will meetfatigue criteria for the empirical case definition.

Discussion

 Reeveset al. (2005) claim that the empirical definition identifies people with CFS ina more precise manner than can occur in the more traditional way of diagnosis.Analyses from this study reveal that the new empirical case definitionidentified 38% of the MDD group as meeting CFS criteria. Cantwell (1996) arguesthat diagnostic criteria should specify which diagnostic instrument to use,what type of informants to interview, and how to determine the presence andseverity of the criteria. The effort by Reeves et al. to specify a certainnumber and type of symptoms that should be present in order to make aparticular diagnosis appears to be overinclusive, particularly for those havinga primarily depressive disorder.

An analysis of the Medical Outcomes SurveyShort-Form-36 illustrates the problems with the cutoff criteria. When using theReeves et al. (2005) cutoff points to classify functional impairment, all threegroups (100%) met criteria for this instrument in Table 3. However, had Reeveset al. selected either Physical Function or Role Physical, betterdifferentiation would have occurred, as there is a significant differencebetween the CFS group and the other two groups for these domains. Becauseindividuals need only to score lower than the 25th percentile in one of thesefour areas in order to meet the CFS criteria, individuals might not have anyreductions in key areas of physical functioning and only impairment in roleemotional areas (e.g., problems with work or other daily activities as a resultof emotional problems).

For Role Emotional, 93% of the MDD/CFS groupand 78% of the MDD group met criteria, a percentage much higher than the CFS group(44%). Ware et al. (2000) found that the mean for Role Emotional for a clinicaldepression group was 38.9, indicating that almost all those with clinicaldepression would meet criteria for being within the lower 25th percentile onthis scale (which was a score of less than or equal to 66.7). In addition, Kingand Jason (2005) compared a group diagnosed with CFS and a group diagnosed withMDD, and the latter group had lower scores than the group with CFS (37.8 vs. 48.9),but both groups would have met the CDC criteria as they both scored below 66.7.In contrast, if the criterion was a score lower than the 25^th percentileon just Physical Function (less than or equal to 70), the participants with CFSwould have met this criterion as their average score was 44, whereas manywithin the MDD group would have not met this criterion as their average scorewas 70.3.

Regarding the Symptom Inventory, 100% of boththe CFS and MDD/CFS groups met criteria, indicating this instrument did notdistinguish the individuals with CFS from individuals with major depression. Itis probable that the Symptom Inventory misclassified the MDD/CFS group forseveral reasons. For example,  theSymptom Inventory asks about the symptom occurrences within the past monthrather than the past 6 months, as required by the Fukuda et al. (1994) casedefinition. The requirement for a participant to report a symptom for 1 monthmight include more individuals within the CFS category (e.g., a person who hasexperienced a physical illness such as influenza or a head cold could very wellhave experienced a severe sore throat for the past month). Even with summedscores for the empirical case definition needing to be greater than or equal to25 (Reeves et al., 2005), the overall level of symptoms might be relatively lowfor patients with classic CFS symptoms (the criterion would be met if anindividual rated only two symptoms as occurring all the time, and one was ofmoderate and the other of severe severity). Similarly, a person with MDD couldendorse symptoms that would easily meet criteria for this scale, such asunrefreshing sleep, impaired memory, and headaches, and muscle pain at amoderate to severe level. However, the most important factor is that theSymptom Inventory does not distinguish critical symptoms for CFS such aspostexertional malaise, unrefreshing sleep, and cognitive difficulties. Eachsymptom is given the same value, which means that a participant reportingsevere and frequent headaches is given the same value as a participant reportingsevere and frequent postexertional malaise. Overall, 14 individuals diagnosedwith MDD scored 25 or higher on the Symptom Inventory and reported four or moresymptoms. This demonstrates that  individualswith primary psychiatric illnesses are not always excluded using the CDCSymptom Inventory. 

The Multidimensional Fatigue Inventory wasused to measure severe fatigue, yet 93% of both the CFS and MDD/CFS groups metcriteria for General Fatigue, while 74% of the MDD group did as well. As forthe criteria that Reeves et al. (2005) used, the primary developer of theMultidimensional Fatigue Inventory had this to say: "Regarding thecriteria suggested by Reeves, we have no paper to back up their decision, butscanning their paper it appears that they used the median of their owndata" (E. M. Smets, personal communication, June 29, 2006). In one studyof three groups with CFS, the mean Multidimensional Fatigue Inventory GeneralFatigue scores were 18.3 to 18.8 (Tiersky, Matheis, DeLuca, Lange, &Nateson, 2003). When assessing Reduced Activity, 85% and 86% of both the CFSand MDD/CFS groups (respectively) met criteria, as did 78% of the MDD group.Therefore, 100% of the CFS and 100% of the MDD/CFS group met the CDC fatiguecriteria. The problem with this instrument is that it is relatively easy tomeet criteria for one of the two categories. In other words, a depressed personcould easily respond positively to questions such as "I get littledone" or "I do very little in a day" and answer negatively to"I feel very active" or "I think I do a lot in a day."Consequently, a depressed person would meet CFS criteria by answering"entirely true" to these types of items.

Inspecting the scores of a person with MDDwho was inappropriately classifiedas having CFS highlights the problems withthe CDC empirical criteria. A 26-year-old female with MDD met criteria for CFSusing the CDC empirical case criteria (Reeves et al., 2005). For the MedicalOutcomes Survey Short-Form-36, she met cutoff points for Social Function (scoring37.5 when needing to score 75) and Role Emotional (scoring 0 when needing toscore66.7). With a clinical diagnosis of MDD, she demonstrated impairment with socialand emotional functioning, two important traits of depression. This personscored 100 on Physical Function, which is the highest possible score on thismeasure, indicating that she had no difficulties with physical functioning,which would be a clear indicator that she did not have CFS. On the CDC SymptomInventory, she reported that postexertional malaise was mild only some of thetime, indicating that she did not have this cardinal symptom of CFS. For thisindividual and others within the MDD/CFS group, the instruments used toidentify cases of CFS did not adequately exclude persons with primarypsychiatric disorders.

Study Limitations

There were biases in using a conveniencesample, and recruitment from a population-based referral source would have beenpreferable, but such samples are expensive to recruit. Also, the sample sizesoverall were relatively small, but even though power was low to detectdifferences, we were able to find a number of significant outcomes, asrepresented in Tables 2 and 3. In addition, we focused on only one psychiatricdisorder, and future studies might include anxiety disorders, which might alsobe misclassified. In addition, there is probably a redundancy in some of ourfindings, as some of the scales are correlated.

There are other ways that might be used todevelop improvements in the CFS case definition. As an example, Jason, Corradi,and Torres-Harding (2007) factor analyzed the core symptoms as defined by theFukuda et al. (1994)  criteria, butthis did not result in interpretable factors. However, when they included alarger group of theoretically defined symptoms in the factor analyses, aninterpretable set of factors did emerge. The following factors were found:neurocognitive (e.g., slowness of thought), vascular (e.g., dizzy afterstanding), inflammation (e.g., chemical sensitivities), muscle/joint (e.g.,pain in multiple joints), infectious (e.g., sore throat), and sleep/postexertional(e.g., unrefreshing sleep). These findings suggest that theoretical andempirical approaches to determining critical symptoms of CFS have considerablemerit. The field of CFS studies needs to be grounded in empirical methods fordetermining a case definition versus more consensus-based efforts.

In conclusion, this study suggests that theReeves et al. (2005) empirical case definition has broadened the criteria suchthat some individuals with a purely psychiatric illness will be inappropriatelydiagnosed as having CFS. The Reeves et al. empirical case definition usedspecific instruments (such as the Medical Outcomes Survey Short-Form-36) tomake diagnostic decisions but included dimensions within them such as roleemotional functioning that were not specific for this illness. Green, Romei,and Natelson (1999) found that 95% of individuals seeking medical treatment forCFS reported feelings of estrangement, and 70% believed that others uniformlyattributed their CFS symptoms to psychological causes. Inappropriate inclusionof pure psychiatric disorders into the CFS samples may further contribute tothe diagnostic skepticism and stigma that individuals with this illness encounter.Several researchers continue to believe that CFS should be considered afunctional somatic syndrome (Barsky & Borus, 1999), characterized bydiffuse, poorly defined symptoms that cause significant subjective distress anddisability and that cannot be corroborated by  consistent documentation of organic pathology. Taylor,Jason, and Schoeny (2001) have challenged this position, but ultimatelyassessment and criteria that fail to capture the unique characteristics ofthese illnesses might inaccurately conclude that only distress and unwellnesscharacterize CFS,thus inappropriately supporting a unitary hypotheticalconstruct called  "functionalsomatic syndrome." Such blurring of diagnostic categories will make iteven more difficult to identify biological markers for this illness, and ifthey are not identified, many scientists will be persuaded that this illness ispsychogenic (Jason & Richman, 2008). Ultimately, using a broad or narrowdefinition of CFS will have important influences on CFS epidemiologic findings,on rates of psychiatric comorbidity, on how patients are treated, andultimately on the likelihood of finding biological markers for this illness.

Note

1. The scale we used had five choices, and weneeded to convert the ratings    to a 4-point scale. We divided the five items by4, which came to .8. We     then made each increment invalue .8.

About the Authors

Leonard A. Jason, PhD, is a professor ofpsychology at DePaul University and the Director of the Center for CommunityResearch. His current interests include myalgic encephalomyelitis/chronicfatigue syndrome, recovery homes, and tobacco control.

Natasha Najar, BA, currently conductsresearch at Northwestern University. She has particular interests in culturalissues.

Nicole Porter, PhD, currently is the projectdirector of a chronic fatigue syndrome (CFS) epidemiologic grant at the Centerfor Community Research, DePaul University. Her interests are in myal- gicencephalomyelitis/CFS, meditation, and dynamic  systems.

Christy Reh, BA, currently is a graduatestudent at the Alder School of Professional Psychology in Chicago, Illinois.

Tables not included in thisposting [Ed]