High Prevalence of Chronic Fatigue in Adult Long-Term Survivors of Acute Lymphoblastic Leukemia and Lymphoma During Childhood and Adolescence

Abstract

Purpose:

Chronic fatigue (CF) is a frequent late effect of cancer, but has been poorly studied in adult survivors of childhood cancer. This cross-sectional study compares the prevalence of CF among adult survivors (childhood leukemia/lymphoma survivors; CLSs) of childhood acute lymphoblastic leukemia (ALL), Hodgkin lymphoma (HL), and non-Hodgkin lymphoma (NHL) to controls, and explored CF's association with disease characteristics, treatment, mental and somatic late effects, and selected markers of low-grade inflammation.

Methods:

A total of 143 male and 147 female CLSs participated in this study, which comprised a questionnaire, clinical examination, and blood samples. The control group consisted of 1405 individuals from the Norwegian general population. The Chalder Fatigue Questionnaire assessed CF.

Results:

Median age at survey and observation time was 29.6 years and 21.1 years, respectively. The prevalence of CF was 27% among CLSs (ALL 22%, NHL 30%, HL 34%) versus 8% among controls. Compared to controls and adjusting for age and sex, the probability of having CF among CLS was elevated (odds ratio [OR]=4.5; 95% confidence interval [CI]: 3.1–6.4). In lymphoma survivors, presence of B-symptoms at diagnosis predicted CF in univariate analysis. Elevated levels of anxiety and depression predicted CF in multivariate analysis. Disease characteristics and treatment or somatic late effects were not associated with CF. Leukocyte, neutrophil, and trombocyte counts were elevated among subjects with CF.

Conclusion:

At a median of 20 years after diagnosis, the prevalence of CF in CLSs is more than three times that of the general population. A persistent low-grade inflammatory response may be involved in the pathogenesis of CF.

Fatigue is a frequent late effect of cancer treatment, and is defined as a persistent, subjective experience of tiredness, exhaustion, and lack of energy that is not relieved by rest.¹ While acute fatigue is a common symptom in all phases of the cancer trajectory, chronic fatigue (CF) is defined as persistent or relapsing fatigue lasting 6 months or longer.² Fatigue negatively impacts cancer survivors' health-related quality of life,³ daily function, and ability to work.⁴

CF has mostly been studied among survivors of adulthood cancer.^5,6 Results from the few studies of adult survivors of childhood cancers including fatigue as a late effect are conflicting, reporting prevalences ranging from 11% to 30%.^7–10 In addition, only one study distinguished between acute and chronic fatigue, and compared the prevalence of CF among childhood cancer survivors to a control group from the general population.¹⁰

Though unknown in detail, the etiology of cancer-related fatigue is viewed to be multifactorial and related to both biological and psychological factors.^1,11 Depression and psychological distress, insomnia, and chronic pain have been found to be associated with CF, but these conditions alone do not explain the prevalence of CF among cancer survivors.¹ Somatic late effects caused by cancer treatment add to the complexity. In survivors of adulthood Hodgkin lymphoma (HL), associations have been reported between CF and pulmonary dysfunction¹² and self-reported cardiac disease.¹³

Studies among survivors of breast and testicular cancer indicate that CF might be associated with a low-grade inflammatory response, reflected by increased levels of C-reactive protein (CRP) and selected cytokines.^14–16 Studying this possible relation may be of particular interest among survivors of lymphogenic malignancies, as B-symptoms (fever, night sweats, weight loss) at diagnosis are associated with increased levels of pro-inflammatory markers.¹⁷ Further, B-symptoms at the time of diagnosis are positively related to CF among long-term HL survivors.^18,19

HL, non-Hodgkin lymphoma (NHL), and acute lymphoblastic leukemia (ALL) represent 50% of all childhood cancers. Today's 5-year survival rates for these lymphogenic malignancies have reached 85%–95%.^20,21 The growing population of these survivors makes studies of late effects, including CF, increasingly important.

We performed a cross-sectional study among adult survivors after childhood ALL, NHL, and HL (adult childhood leukemia/lymphoma survivors; CLSs) with the following aims: (1) to determine the prevalence of CF among the CLSs compared to controls from the general population, and (2) to explore factors associated with CF, including mental and somatic late effects and markers of low-grade inflammation.

Based on the findings from HL survivors treated in adulthood, it was hypothesized that the prevalence of CF would be higher among CLSs than in the general population. In addition, the prevalence of CF in HL survivors was expected to be higher than among ALL and NHL survivors. We expected that mental and somatic late effects would be associated with CLSs' report of CF. Furthermore, it was hypothesized that a low-grade inflammatory condition—reflected as elevated levels of CRP, leukocytes, and trombocytes—would be associated with CF.

Methods

Patients and study design

Eligible patients were identified by the Cancer Registry of Norway with the following eligibility criteria: (1) treatment for HL as their first cancer at university hospitals in Norway or treatment at Oslo University Hospital (OUH) for ALL or NHL; (2) diagnosed between 1970 and 2000 (1970 and 2002 for ALL due to later point of time of inclusion); (3) age at diagnosis ≤18 years (≤16 years for ALL, as only pediatric patients were included) and survival for ≥5 years; (4) age at survey >18 years; (5) alive as of June 2007 (April 2009 for ALL).

For HL, the survey was national. Due to resource limitations, the survey was restricted to survivors of ALL and NHL from the South-Eastern Health Region of Norway, covering approximately 50% of the Norwegian population.

Participants with a second cancer or ongoing pregnancy were excluded from all analyses concerning intergroup comparisons between HL, ALL, and NHL survivors. Such survivors were, however, included in analyses that compared survivors to controls, as the participants of the control group were unselected in terms of these conditions.

From 2008 to 2011, eligible CLSs were invited to participate in a cross-sectional study on late effects. The survivors were invited to a general health check-up without specifically mentioning CF as a possible outcome. The study included a mailed questionnaire and a 2-day outpatient examination at OUH, which included a clinical examination, echocardiography, lung function tests, and fasting blood sampling. Levels of thyroid-stimulating hormone (TSH), thyroxin (T4), B-type natriuretic peptide (proBNP), and CRP, and leukocyte and trombocyte counts were measured. The lowest detection level of CRP was 0.6 mg/L. A CRP level >10 mg/L was considered to reflect possible acute infections, and survivors with CRP levels >10 mg/L (n=14) were excluded from the analyses of the association between CF and inflammatory markers. They were, however, included in the other analyses, as the questionnaire had been filled in some weeks prior to the 2-day visit.

The date of study was defined as the date of clinical examination (if present) or the date that the questionnaire was completed. Observation time was defined as the period from the date of diagnosis to the date of study. Information about prior treatment and clinical course was retrieved from medical records. The initial extent of the lymphomas were classified according to the Ann Arbor classification.²² B-symptoms at diagnosis comprised a fever >38°C and/or night sweats for more than 2 weeks preceding diagnosis and/or >10% weight loss during the last 6 months.

Treatment

Treatment of ALL was predominantly based on chemotherapy alone and is described elsewhere.^23,24 Treatment for the lymphoma survivor participants has been previously summarized,²⁵ and in most cases included a combination of chemotherapy and radiotherapy, with large-field radiotherapy applied to patients with HL in the 1970s. Treatment periods were defined as before 1986 or later.

Clinical findings

Echocardiography was performed (GE Vivid 7 or E9; GE Vingmed, Horten, Norway) and analyzed following standard recommendations.²⁶ Aortic valve dysfunction (aortic valve stenosis and/or regurgitation) was identified by echocardiography using two-dimensional imaging. Doppler and color Doppler data were categorized as present when a survivor obtained a score of ≥2 on a four-grade scale (i.e., valvular dysfunction: 0=normal, 1=small, 2=moderate, and 3=severe). Left ventricular shortening fraction (SF)≤28% defined reduced systolic function. ProBNP≥20 pmol/L was considered to reflect left ventricular dysfunction. Reduced cardiac function was defined as the presence of aortic valve dysfunction and/or SF≤28% and/or proBNP≥20 pmol/L. Total lung capacity (TLC)≤80% of predicted level was defined as lung volume reduction. Gas transfer impairment was defined as diffusion capacity of the lung for carbon monoxide (DLCO) of ≤80% of predicted level. Reduced lung function was defined as lung volume reduction and/or gas transfer impairment. TSH>3.4 mIU/L and/or T4<9.0 pmol/L was defined as unsubstituted ongoing hypothyroidism.

Questionnaires

The mailed questionnaire included the 11-item Chalder Fatigue Questionnaire (FQ), widely used for assessment of fatigue severity and for case detection in clinical and epidemiological studies.²⁷ Initial Likert scoring (0, 1, 2, 3) defines total fatigue (all 11 items) by simple addition with higher scores implying higher levels of fatigue. Two additional items ask for the duration and extent of fatigue. The scores (0, 1, 2, 3) are dichotomized (0, 0, 1, 1) for the definition of CF. CF is defined by a sum score of ≥4 for all 11 dichotomized items and a duration of ≥6 months.^27,28

Mental distress was assessed by the Hospital Anxiety and Depression Scale (HADS) consisting of 14 items—7 comprising the depression subscale and 7 comprising the anxiety subscale. Each item is scored from 0 (minimum) to 3 (maximum). The HADS total score is the sum scores for all 14 items.

The questionnaire also asked for partnership status and highest level of education, dichotomized as ≤11 years and >12 years.

Controls

Of 3254 subjects representative of the entire Norwegian population and previously included in a survey assessing CF in 1996,²⁹ 1405 were between 19 and 50 years old and served as controls in the present study.

Statistics

Standard descriptive analyses were applied using SPSS version 18 with χ² tests, t-tests, and Mann–Whitney U tests analyzing differences between groups. The Pearson product–moment correlation coefficient assessed correlations. Predictors of CF were evaluated by logistic regression analysis (odds ratios [OR] and 95% confidence intervals [CI]). Variables with a significance level of p<0.1 in the univariate logistic regression analyses were included in the multivariate analyses. The significance level was set at p≤0.05, two-sided.

Age at survey, sex, educational level, and partnership status were adjusted for when the CLSs were compared to the controls.

The ALL and the lymphoma samples differed in terms of ages at diagnosis and types of treatment and burden. Extent of disease was not categorized in ALL patients. When exploring effects of disease characteristics and treatment- and somatic late effects on CF, ALL and lymphoma survivors were analyzed separately, with NHL and HL survivors combined.

Ethics

The study was approved by Regional Committee for Medical Research Ethics.

Results

Patient characteristics

Of 430 eligible survivors invited, 300 responded (70%). A total of 21 survivors were excluded (10 had incomplete questionnaires, 6 were pregnant, and 5 had a second cancer diagnosis), leaving 279 (65%) for studying the prevalence of CF. Of these, 246 attended the clinical examination.

The 134 non-responders were more likely to be male (64%, p=0.001) or lymphoma survivors (59%, p=0.03), with no statistically significant differences as to age at diagnosis or age at survey or follow-up (data not shown). Further analysis of non-responders was not feasible due to restrictions from the ethical committee.

Demographics of the CLSs and controls are summarized in Table 1. Disease characteristics and treatments of the CLSs are summarized in Table 2. The median age at diagnosis for the CLSs was 9.5 years (range=0.3–18.4), with a median observation time of 21.1 years (range=6.9–39.4). The median age at survey was 29.6 years (range=18.3–54.5) for the CLSs, compared to 34.0 years (range=19.0–50.0) among the controls (p<0.001). Compared to lymphoma survivors, ALL survivors were younger at both diagnosis and study (p<0.001) but with a longer observation time (p=0.005).

Table 1.

Demographics for Childhood Leukemia/Lymphoma Survivors and Controls

	Total CLSs N=290	ALL n=151	NHL n=47	HL n=92	Controls N=1405	p ^a
Age at survey, years (range)	29.6 (18.3–54.5)	28.0 (18.3–45.7)	29.3 (18.6–48.2)	35.0 (19.2–54.5)	34.0 (19.0–50.0)	<0.001
Female	139 (50%)	78 (47%)	16 (65%)	45 (53%)	720 (51%)	<0.001
Male	140 (50%)	70 (53%)	30 (36%)	40 (47%)	703 (49%)	<0.001
Education >12 years	175 (63%)	91 (62%)	22 (48%)	62 (73%)	1215 (85%)	<0.001
In a partnership	145 (52%)	69 (47%)	24 (53%)	52 (61%)	929 (65%)	<0.001
Age at diagnosis, years (range)	9.5 (0.3–18.4)	4.8 (0.3–16.0)	12.2 (4.1–17.8)	14.6 (2.2–18.4)
Observation time, years (range)	21.1 (6.9–39.4)	22.1 (7.1–39.4)	18.4 (8.4–36.0)	20.6 (6.9–37.0)

CLSs versus controls.

ALL, acute lymphoblastic leukemia; CLSs, childhood leukemia/lymphoma survivors; HL, Hodgkin lymphoma; NHL, non-Hodgkin lymphoma.

Table 2.

Disease Characteristics, Treatment and Clinical Findings of Childhood Leukemia/Lymphoma Survivors

	Total	ALL	NHL	HL
Disease characteristics	N=279	n=148	n=46	n=85
Stage I–II		NA	24 (52%)	60 (71%)
Stage III–IV		NA	22 (48%)	25 (29%)
B-symptoms^a
No		NA	17 (37%)	58 (68%)
Yes		NA	11 (24%)	18 (21%)
Unknown		NA	18 (38%)	9 (11%)
Treatment	N=279	n=148	n=46	n=85
Radiotherapy only	13 (5%)	0	1 (2%)	12 (14%)
Chemotherapy only	179 (64%)	132 (89%)	27 (59%)	20 (23%)
Radiotherapy + chemotherapy	87 (31%)	16 (11%)	18 (39%)	53 (62%)
Radiotherapy, CNS	19 (7%)	16^b (11%)	3 (7%)	0
Dose (Gy, median [range])^c	20 (12–54)	20 (12–25)	(24–54)	–
Radiotherapy, mediastinum	69 (25%)	0	7 (15%)	62 (73%)
Dose (Gy, median [range])^c	36 (18–44)	–	40 (22–40)	36 (18–44)
Radiotherapy, other	12 (4%)	0	9^d (20%)	3 (4%)
Dose (Gy, median [range])^c	40 (13–41)	–	27 (13–41.4)	40 (30–41.4)
AC	199 (71%)	104 (70%)	42 (91%)	53 (62%)
Dose (mg/m², median [range])^c	150 (40–510)	120 (40–510)	150 (50–460)	160 (50–401)
HDT	12 (4%)	3 (2%)	5 (11%)	4 (5%)
Relapse	25 (9%)	18 (12%)	3 (7%)	4 (5%)
Clinical findings ^e	N=246^e	n=130^e	n=42^e	n=74^e
Impaired heart function	73 (28%)	35 (27%)	5 (12%)	33 (45%)
Reduced SF	55 (22%)	33 (25%)	5 (12%)	17 (23%)
ProBNP ≥20 pmol/L	21 (9%)	4 (3%)	1 (2%)	16 (22%)
Aortic valve dysfunction	16 (7%)	0	0	16 (22%)
Reduced lung function	70 (28%)	35 (27%)	10 (24%)	25 (34%)
Lung volume reduction	19 (8%)	7 (5%)	3 (7%)	9 (12%)
Impaired gas transfer	60 (24%)	33 (25%)	7 (17%)	20 (27%)
BMI>30 kg/m²	31 (12%)	22 (17%)	4 (10%)	5 (7%)
Present hypothyroidism	49^f (20%)	16 (12%)	5 (12%)	28 (38%)
HADS total score^g (mean±SD)	8.4±6.2	8.7±6.6	7.9±5.2	8.1±6.0

Includes night sweats, fever of unknown cause, and weight loss at time of diagnosis.

Including nine irradiated to medulla spinalis.

Per patient receiving the treatment.

Including five with total body irradiation.

33 patients did not perform a clinical examination (questionnaire only).

48 had elevated TSH only; one had elevated T4 and TSH.

Based on questionnaire: N=279; n_ALL=148, n_NHL=46, n_HL=85.

ALL, acute lymphoblastic leukemia; AC, anthracycline; BMI, body mass index; CNS, central nervous system; HADS, Hospital Anxiety and Depression Scale; HDT, high-dose chemotherapy with stem cell support; HL, Hodgkin lymphoma; NA, not applicable; NHL, non-Hodgkin lymphoma; proBNP, B-type natriuretic peptide; SD, standard deviation; SF, shortening fraction of left ventricle; T4, thyroxin; TSH, thyroid-stimulating hormone.

Prevalence of CF among CLSs versus controls

Twenty-eight percent of the CLSs had CF versus 8% of the controls (p<0.001; Table 3). When adjusting for sex and age at survey, the probability of having CF was 4.5 times higher among survivors compared with controls (OR=4.5; 95% CI: 3.1–6.4). This probability was highest among the HL survivors (OR=5.9; 95% CI: 3.6–9.7), followed by survivors of NHL (OR=4.4; 95% CI: 2.2–9.0) and ALL (OR=3.6; 95% CI: 2.3–5.7). Impaired heart function and reduced lung function were most frequently observed in HL survivors.

Table 3.

Chronic Fatigue Among Childhood Leukemia/Lymphoma Survivors (N=290) and Controls (N=1405)

	Crude			Adjusted ^a
	CF/total (%)	OR (95% CI)	p	OR (95% CI)	p
Controls	118/1405 (8%)	ref		ref
CLSs	79/290 (28%)	4.2 (3.0–5.7)	<0.001	4.5 (3.1–6.4)	<0.001
Controls	118/1405 (8%)	ref		ref
CLS type:
ALL	34/151 (23%)	3.2 (2.1–4.9)	<0.001	3.6 (2.3–5.7)	<0.001
NHL	14/47 (30%)	4.6 (2.1–8.9)	<0.001	4.4 (2.2–9.0)	<0.001
HL	32/92 (35%)	5.8 (3.6–9.3)	<0.001	5.9 (3.6–9.7)	<0.001

Note: CF defined as a Chalder Fatigue Questionnaire sum score of ≥4 for all 11 dichotomized items and a duration of ≥6 months.

Adjusted for age at survey, sex, education, and partnership status.

ALL, acute lymphoblastic leukemia; CI, confidence interval; CF, chronic fatigue; CLSs, childhood leukemia/lymphoma survivors; HL, Hodgkin lymphoma; NHL, non-Hodgkin lymphoma; OR, odds ratio; ref, reference.

Factors associated with CF among the CLSs

Compared to ALL survivors, HL survivors displayed a significantly increased probability of reporting CF, but this was not observed in NHL survivors (Table 4). The chance of having CF increased with age at survey and with the HADS total score. Treatment period, sex, partnership status, and educational level did not influence CLSs' probability of having CF. In the multivariate analysis, the elevated probability of having CF among HL survivors and the age-related association disappeared, with only the HADS total score remaining significantly associated with CF (OR=1.15; 95% CI: 1.1–1.2).

Table 4.

Risk of Chronic Fatigue Among Childhood Leukemia/Lymphoma Survivors (N=279) Related to Diagnosis, Demographic Factors, BMI, Thyroid Function, and HADS

	Univariate		Multivariate
	OR (95% CI)	p	OR (95% CI)	p
ALL	ref		ref
NHL	1.4 (0.6–2.9)	0.4	1.5 (0.6–3.4)	0.4
HL	1.8 (1.0–3.3)	0.05	1.7 (0.8–3.5)	0.2
Age at survey	1.04 (1.00–1.07)	0.01	1.05 (1.0–1.1)	0.1
Treatment after 1985	ref		ref
Treatment 1970–1985	1.3 (0.8–2.2)	0.3	0.8 (0.3–2.1)	0.7
Male	ref		ref
Female	0.9 (0.5–1.5)	0.7	0.8 (0.46–1.5)	0.6
In a partnership	ref
Not in a partnership	0.7 0.4–1.3	0.3
Education≥12 years	ref
Education ≤11 years	1.6 (0.8–2.7)	0.1
BMI<30 kg/m²	ref
BMI≥30 kg/m²	1.8 (0.8–4.0)	0.1
Thyroid status normal	ref		ref
Present hypothyroidism	1.8 (0.9–3.4)	0.09	1.4 (0.7–3.0)	0.4
HADS total score	1.1 (1.1–1.2)	<0.001	1.15 (1.1–1.2)	<0.001

Note: CF defined as a Chalder Fatigue Questionnaire sum score ≥4 for all 11 dichotomized items and a duration of ≥6 months. Only variables with significance level of p<0.1 in the univariate logistic regression analyses were included in the multivariate analyses.

ALL, acute lymphoblastic leukemia; BMI, body mass index; CI, confidence interval; CLSs, childhood leukemia/lymphoma survivors; CF, chronic fatigue; HADS, Hospital Anxiety and Depression Scale; HL, Hodgkin lymphoma; NHL, non-Hodgkin lymphoma; OR, odds ratio; ref, reference.

Further sub-analyses separated ALL survivors from HL/NHL survivors (Table 5). In multivariate analysis, ALL was the only group for which increasing age at survey was associated with increased probability of reporting CF (OR=1.1; 95% CI: 1.0–1.2). No significant associations were found between CF and treatment period, sex, relapse, radiotherapy, cumulative dose of anthracycline, reduced cardiac function, or reduced lung function in either of the diagnostic groups. The presence of B-symptoms in lymphoma survivors was significantly associated with increased probability of CF (OR=2.5; 95% CI: 1.0–6.2).

Table 5.

Risk for Chronic Fatigue Related to Disease Characteristics, Treatment Factors, and Heart and Lung Function for Survivors of Childhood Acute Lymphoblastic Leukemia and Lymphoma

	Univariate		Multivariate
	OR (95% CI)	p	OR (95% CI)	p
ALL (n=148)
Age at survey	1.0 (1.0–1.1)	0.003	1.1 (1.0–1.2)	0.01
Treatment after 1985	ref		ref
Treatment 1970–1985	2.0 (0.9–4.4)	0.08	0.6 (0.2–2.1)	0.4
Male	ref		ref
Female	0.8 (0.4–1.7)	0.6	0.9 (0.4–2.1)	0.8
No relapse	ref
Relapse	0.7 (0.2–2.5)	0.5
AC, cumulative dose	1.0 (1.0–1.0)	0.9
No RT	ref
RT CNS	1.2 (0.3–3.9)	0.8
Normal heart function	ref
Reduced heart function	1.7 (0.7–4.1)	0.2
Normal lung function	ref
Reduced lung function	1.0 (0.4–2.9)	0.8
Lymphoma (n=131)
Age at survey	1.0 (1.0–1.1)	0.9	1.0 (0.9–1.1)	0.5
Treatment 1970–1985	ref		ref
Treatment after 1985	0.9 (0.4–1.9)	0.8	0.6 (0.2–2.3)	0.5
Male	ref		ref
Female	1.1 (0.5–2.2)	0.9	0.9 (0.4–2.0)	0.9
No relapse	ref
Relapse	0.8 (0.3–2.2)	0.8
Stage I–II	ref
Stage III–IV	1.4 (0.6–2.9)	0.4
B-symptoms^a
No	ref		ref
Yes	2.4 (1.0–5.8)	0.05	2.5 (1.0–6.2)	0.05
Unknown	0.9 (0.3–2.4)	0.8	1.0 (0.3–2.7)	0.9
AC, cumulative dose	1.0 (1.0–1.0)	0.8
No RT or “Other”	ref
RT mediastinum	1.0 (0.5–2.1)	0.9
Normal heart function	ref
Reduced heart function	1.6 (0.7–3.7)	0.2
Normal lung function	ref
Reduced lung function	1.2 (0.5–2.9)	0.7

Note: CF defined as a Chalder Fatigue Questionnaire sum score of ≥4 for all 11 dichotomized items and a duration of ≥6 months. Only variables with significance level p<0.1 in the univariate logistic regression analyses were included in the multivariate analyses.

Includes night sweats, fever of unknown cause, and weight loss at time of diagnosis.

ALL, acute lymphoblastic leukemia; AC, anthracycline; CI, confidence interval; CF, chronic fatigue; CNS, central nervous system; OR, odds ratio; ref, reference; RT, radiotherapy.

The Pearson product–moment correlation coefficient between the total fatigue score and the HADS score was 0.54. Leukocyte, neutrophil, and trombocyte counts were significantly higher among survivors with CF compared to those without CF (Table 6), though with minimal numeric differences between CLSs with or without CF, these variables were not included in the regression analyses.

Table 6.

Inflammatory Markers Related to Chronic Fatigue in Childhood Leukemia/Lymphoma Survivors (N=230^a)

	No CF (n=164)	CF (n=66)	p
CRP (mg/L) (median [range])	1.0 (<0.6–10.0)	1.0 (<0.6–8.3)	0.9^b
Leukocytes (×10⁹) (mean±SD)	5.9±1.6	6.5±2.2	0.04^c
Neutrophils (×10⁹) (mean±SD)	3.0±1.0	3.4±1.5	0.02^c
Lymphocytes (×10⁹) (mean±SD)	2.1±0.9	2.2±1.0	0.4^c
Trombocytes (×10⁹) (mean±SD)	281±81	306±84	0.04^c

Note: CF defined as a Chalder Fatigue Questionnaire sum score of ≥4 for all 11 dichotomized items and a duration of ≥6 months.

Two out-clinic patients missed blood tests; 14 patients excluded due to CRP>10 mg/L. ^bMann–Whitney U test. ^ct-test.

CF, chronic fatigue; CLSs, childhood leukemia/lymphoma survivors; CRP, C-reactive protein; SD, standard deviation.

Discussion

In this population-based study of adult CLSs with a median observation time of 20 years, the prevalence of CF was 3.5 times higher than in the general population. Higher age at survey was associated with CF among the ALL survivors, whereas the presence of B-symptoms at diagnosis increased the probability of CF among the lymphoma survivors. Leukocyte, neutrophil, and trombocyte counts were significantly higher among CLSs with CF compared to those without. Neither previous cancer treatment nor somatic clinical findings impacted the probability of having CF.

In spite of the relatively young age at survey (median=29.6 years) and a very long follow-up time (median=21.1 years), the prevalence of CF among CLSs (all diagnoses included) in this study was as high as previously reported among survivors of adult HL.²⁸ To our surprise, the prevalence of CF was not lower among the ALL survivors compared to the lymphoma survivors after adjusting for age, sex, and treatment period. Three other studies have explored fatigue in adult survivors of childhood ALL and lymphoma, all with comparable observation times (14–23 years).^7–9 Meeske et al. reported a 30% prevalence of fatigue in survivors of childhood ALL, but the response rate was low (44%) and no control group was included.⁸ Mulrooney et al. found a prevalence of 15% in ALL survivors compared to 10% in sibling controls. This study also demonstrated a prevalence of 21% in childhood HL survivors.⁷ Langeveld et al. were unable to demonstrate an increased risk of fatigue for either childhood ALL or NHL survivors compared to a control group,⁹ though the validity of the instrument used in that study (the Multidimensional Fatigue Inventory [MFI-20]) has been questioned for use among cancer patients.³⁰ Further, direct comparisons across these studies and ours are hampered by the use of four different instruments for assessment of fatigue (Piper Fatigue Scale, Functional Assessment of Chronic Illness Therapy-Fatigue [FACIT-F], MFI-20, and FQ). Furthermore, the three latter studies assessed acute fatigue at time of survey while we, on the contrary, have integrated a duration of at least 6 months in our definition of CF. Jóhannsdóttir et al. reported a prevalence of CF after childhood cancer almost double that of the Norwegian general population (11% versus 6%),¹⁰ but the study consisted of survivors of rare types of pediatric cancers (acute myeloid leukemia, infratentorial astrocytoma, and Wilms' tumor) and had a relatively young median age at survey (24 years).

A study of adult HL survivors found an association between impaired DLCO and CF, but applied a somewhat stricter definition of reduced lung function (DLCO ≤75% of predicted compared to ≤80% in the present study).¹² Mulrooney et al.'s much larger study sample of questionnaires from 2600 adult childhood cancer survivors demonstrated an association between fatigue and pulmonary fibrosis, and also between fatigue and congestive heart failure.⁷ In the present study, no associations emerged between CF and reduced lung or heart function. However, the findings in Mulrooney et al.'s study were based on patients' reports of these late effects, whereas we assessed heart and pulmonary functions by objective tests. One explanation for the lack of associations between somatic late effects and CF in our study may be that the assessed cardiac and pulmonary findings generally were at a subclinical state and thus probably not symptomatic at the time of study.

The association between fatigue and psychological distress is well documented.^1,31 The Pearson product–moment correlation coefficient between the total fatigue score and the HADS total score in our material was 0.54, confirming the expected moderate correlation between the two scales.

Among the ALL survivors, age was the only factor associated with CF. This result was not observed in the lymphoma group consisting of somewhat older participants. However, among survivors aged >30 years at survey, the prevalence was similar in all survivors (ALL 34% versus lymphoma 30%, p=0.6). The question is open as to whether at least some CLSs are particularly susceptible to the physiological age-dependent increase of CF. In the general population, the prevalence of fatigue has been demonstrated to increase with age.³² Still, the present sample's size limits the possibility of drawing firm conclusions as to differences in CF's relation to age.

As demonstrated previously among adult HL survivors,²⁸ lymphoma survivors in this study that presented with B-symptoms at diagnosis had an elevated probability of having CF, although the groups were small, with the risk of type I error. As B-symptoms at diagnosis are associated with an increase in inflammatory markers,¹⁷ our finding supports the hypothesis that the inflammatory system may be involved in the mechanism of CF, at least after malignant lymphoma. The significantly (though minimally) elevated blood cell counts among the survivors with CF as compared to those without CF seem to support our hypothesis, and these serological findings are in concordance with other studies.^14–16

Our population-based study, including a control group from the Norwegian general population, studied the impact of objectively assessed late effects upon CF. A response rate of 70% must be considered high in this patient population with a follow-up period of 20 years.³³ Whether survivors with CF are overrepresented or underrepresented in our study cannot be determined. A previous study of survivors of adult HL suggested that non-participants displayed more comorbidity than participants. Hence, more non-participants may suffer from fatigue, which possibly limits their survey participation.³⁴

The heterogeneity between the three diagnostic groups (in terms of treatment burden and modalities, and different ages at diagnosis) implied that disease characteristics and treatment-related variables had to be analyzed by diagnostic subgroups. Small group sizes may have limited the interpretation of the intergroup differences, thereby also increasing the risk of type II errors. Finally, this study primarily dealt with the relationship between CF and survivors' somatic health status, but it should be noted that other conditions such as pain, sleep disturbances, and physical inactivity have also shown associations with cancer-related fatigue.^7,35

Conclusion

We found a 28% prevalence of CF among adult CLSs as compared to 8% in controls. The etiology of CF remains unclear; in our study, objectively measured subclinical cardiac and pulmonary sequelae after cancer treatment were not associated with CF. The association between and CF and selected inflammatory markers lends some support to the hypothesis that a low-grade inflammatory response may be involved in the pathogenesis of CF in CLSs.

Footnotes

Acknowledgments

We are thankful for practical help from Siri L. Hess. Eva Widing, Unni Blom, Svein Kolmannsskog, Trond Flægstad, and Geir Tjønnfjord were helpful in the eligibility process. Richard Massey performed the echocardiograms. This work was supported by “Helse Sørøst HF.”

Disclaimer

Results from this study were previously presented at the 2011 European Symposium on Late Complications after Childhood Cancer 2011.³⁶

Author Disclosure Statement

No competing financial interests exist.

References

Stone

, Minton

. Cancer-related fatigue. Eur J Cancer., 2008; 44,8:1097–104.

Fukuda

, Straus

, Hickie

et al. The chronic fatigue syndrome: a comprehensive approach to its definition and study. International Chronic Fatigue Syndrome Study Group. Ann Intern Med., 1994; 121,12:953–9.

Hjermstad

, Oldervoll

, Fossa

et al. Quality of life in long-term Hodgkin's disease survivors with chronic fatigue. Eur J Cancer, 2006; 42,3:327–33.

Spelten

, Verbeek

, Uitterhoeve

et al. Cancer, fatigue and the return of patients to work-a prospective cohort study. Eur J Cancer, 2003; 39,11:1562–7.

Ganz

, Bower

. Cancer related fatigue: a focus on breast cancer and Hodgkin's disease survivors. Acta Oncol., 2007; 46,4:474–9.

Orre

, Fossa

, Murison

et al. Chronic cancer-related fatigue in long-term survivors of testicular cancer. J Psychosom Res., 2008; 64,4:363–71.

Mulrooney

, Ness

, Neglia

et al.

Fatigue and sleep disturbance in adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study (CCSS)

Sleep, 2008; 31,2:271–81.

Meeske

, Siegel

, Globe

et al. Prevalence and correlates of fatigue in long-term survivors of childhood leukemia. J Clin Oncol., 2005; 23,24:5501–10.

Langeveld

, Grootenhuis

, Voute

et al. No excess fatigue in young adult survivors of childhood cancer. Eur J Cancer, 2003; 39,2:204–14.

10.

Jóhannsdóttir

, Hjermstad

, Moum

et al. Increased prevalence of chronic fatigue among survivors of childhood cancers: a population-based study. Pediatr Blood Cancer, 2012; 58,3:415–20.

11.

Prue

, Rankin

, Allen

et al. Cancer-related fatigue: a critical appraisal. Eur J Cancer, 2006; 42,7:846–63.

12.

Knobel

, Havard

, Brit

et al. Late medical complications and fatigue in Hodgkin's disease survivors. J Clin Oncol., 2001; 19,13:3226–33.

13.

, Li

, Recklitis

et al. A comparison between long-term survivors of Hodgkin's disease and their siblings on fatigue level and factors predicting for increased fatigue. Ann Oncol., 2005; 16,12:1949–55.

14.

Orre

, Murison

, Dahl

et al. Levels of circulating interleukin-1 receptor antagonist and C-reactive protein in long-term survivors of testicular cancer with chronic cancer-related fatigue. Brain Behav Immun., 2009; 23,6:868–74.

15.

Bower

. Cancer-related fatigue: links with inflammation in cancer patients and survivors. Brain Behav Immun., 2007; 21,7:863–71.

16.

Alexander

, Minton

, Andrews

et al. A comparison of the characteristics of disease-free breast cancer survivors with or without cancer-related fatigue syndrome. Eur J Cancer, 2009; 45,3:384–92.

17.

Seymour

, Talpaz

, Cabanillas

et al. Serum interleukin-6 levels correlate with prognosis in diffuse large-cell lymphoma. J Clin Oncol., 1995; 13,3:575–82.

18.

Hjermstad

, Fossa

, Oldervoll

et al. Fatigue in long-term Hodgkin's Disease survivors: a follow-up study. J Clin Oncol., 2005; 23,27:6587–95.

19.

Ruffer

, Flechtner

, Tralls

et al.

Fatigue in long-term survivors of Hodgkin's lymphoma: a report from the German Hodgkin Lymphoma Study Group (GHSG)

Eur J Cancer, 2003; 39,15:2179–86.

20.

Kaatsch

. Epidemiology of childhood cancer. Cancer Treat Rev., 2010; 36,4:277–85.

21.

Gatta

, Corazziari

, Magnani

et al. Childhood cancer survival in Europe. Ann Oncol., 2003; 14,Suppl 5:v119–27.

22.

Lister

, Crowther

, Sutcliffe

et al. Report of a committee convened to discuss the evaluation and staging of patients with Hodgkin's disease: Cotswolds meeting. J Clin Oncol., 1989; 7,11:1630–6.

23.

Moe

, Seip

, Finne

. Intermediate dose methotrexate (IDM) in childhood acute lymphocytic leukemia in Norway. Preliminary results of a national treatment program. Acta Paediatr Scand., 1981; 70,1:73–9.

24.

Gustafsson

, Schmiegelow

, Forestier

et al.

Improving outcome through two decades in childhood ALL in the Nordic countries: the impact of high-dose methotrexate in the reduction of CNS irradiation. Nordic Society of Pediatric Haematology and Oncology (NOPHO)

Leukemia., 2000; 14,12:2267–75.

25.

Hamre

, Kiserud

, Ruud

et al. Gonadal function and parenthood 20 years after treatment for childhood lymphoma: a cross-sectional study. Pediatr Blood Cancer, 2012; 59,2:271–7.

26.

Lang

, Bierig

, Devereux

et al. Recommendations for chamber quantification. J Am Soc Echocardiogr., 2005; 18,12:1440–63.

27.

Chalder

, Berelowitz

, Pawlikowska

et al. Development of a fatigue scale. J Psychosom Res., 1993; 37,2:147–53.

28.

Loge

, Abrahamsen

, Ekeberg

, Kaasa

. Hodgkin's disease survivors more fatigued than the general population. J Clin Oncol., 1999; 17,1:253–61.

29.

Loge

, Ekeberg

, Kaasa

. Fatigue in the general Norwegian population: normative data and associations. J Psychosom Res., 1998; 45,1:53–65.

30.

Minton

, Stone

. A systematic review of the scales used for the measurement of cancer-related fatigue (CRF) Ann Oncol., 2009; 20,1:17–25.

31.

Loge

, Abrahamsen

, Ekeberg , Kaasa

. Fatigue and psychiatric morbidity among Hodgkin's disease survivors. J Pain Symptom Manage, 2000; 19,2:91–9.

32.

Schwarz

, Krauss

, Hinz

. Fatigue in the general population. Onkologie., 2003; 26,2:140–4.

33.

Asch

, Jedrziewski

, Christakis

. Response rates to mail surveys published in medical journals. J Clin Epidemiol., 1997; 50,10:1129–36.

34.

Kiserud

, Loge

, Fossa

et al. Mortality is persistently increased in Hodgkin's lymphoma survivors. Eur J Cancer, 2010; 46,9:1632–9.

35.

Berger

, Gerber

, Mayer

. Cancer-related fatigue: implications for breast cancer survivors. Cancer, 2012; 118,8 Suppl:2261–9.

36.

Zeller

, Hamre

, Kanelloupolos

et al. High prevalence of chronic fatigue in long term survivors of childhood acute lymphoblastic leukemia and lymphoma. A population based study. Abstract presented at European Symposium on Late Complications After Childhood Cancer, Amsterdam, The Netherlands, September 29–30, 2011.