Functional Somatic Syndromes and Childhood Physical Abuse in Women: Data From a Representative Community-Based Sample

Abstract

This study investigated whether childhood physical abuse was associated with functional somatic syndromes (FSS) in women while controlling for age, race, and four clusters of potentially confounding factors: (a) Other childhood adversities, (b) adult health behaviors, (c) socioeconomic status and stressors, and (d) mental health. A regional subsample of the 2005 Canadian Community Health Survey of 7,342 women was used. Women reported whether they had been diagnosed with chronic fatigue syndrome (CFS), fibromyalgia (Fm), irritable bowel syndrome (IBS), or multiple chemical sensitivities (MCS). Fully 749 reported having been physically abused by someone close to them during their youth. When controlling for potentially confounding factors, childhood physical abuse was significantly associated with CFS (OR = 2.11; 95% CI = 1.22, 3.65), Fm (OR = 1.65; 95% CI = 1.08, 2.52), and MCS (OR = 2.82; 95% = CI 1.90, 4.17). Clinicians using reattribution and stepped care approaches in the management of FSS should assess for a history of abuse.

Keywords:

• child abuse
• chronic fatigue
• fibromyalgia
• functional somatic syndrome
• idiopathic environmental intolerance
• irritable bowel
• multiple chemical sensitivities
• reattribution
• stepped care approaches

Mental health professionals, primary care medical doctors, and medical specialists frequently treat clients with functional somatic syndromes (FSS). FSS have been defined as patterns of somatic symptoms “for which adequate examination does not reveal sufficiently explanatory structural or other specified pathology” (Henningsen, Zipfel, & Herzog, 2007, p. 946); syndromes that are associated with negative outcomes that are not explained by another medical or psychiatric disorder (Kroenke, Sharpe, & Sykes, 2007; Wessely & White, 2004); and, “conditions where the patient complains of physical symptoms that cause excessive worry or discomfort, or lead the patient to seek treatment but for which no adequate organ pathology or patho-physiological basis can be found” (Fink, Rosendal, & Toft, 2002, p. 99). This study focuses on four types of FSS including (a) chronic fatigue syndrome (CFS), (b) fibromyalgia (Fm), (c) irritable bowel syndrome (IBS), and (d) multiple chemical sensitivities (MCS), which is also known as idiopathic environmental intolerance (IEI).

CFS is characterized by severe fatigue and a combination of symptoms that most often include impairment in concentration and short-term memory, sleep disturbances, and musculoskeletal pain (Fukuda et al., 1994). Epidemiological research has indicated that CFS is prevalent in 0.24% to 2.54% of the population (Reeves et al., 2007; Reyes et al., 2003). Fm is diagnosed by at least 3 months of widespread pain throughout at least 11 of 18 identified tender points (Wolfe, Ross, Anderson, Russell, & Hebert, 1995). Using these criteria, Fm has been found to occur in 2% to 4% of the population (Chakrabarty & Zoorob, 2007; Wolfe et al., 1995). IBS is characterized by abdominal pain or discomfort that is linked with disordered defecation. IBS is the most common of the four FSS; using common diagnostic criteria, it has been found to occur in 15% of the population (Drossman, Camilleri, Mayer, & Whitehead, 2002). Finally, MCS represent a hypersensitivity to common chemical products such as perfume, fresh paint, and pesticides. Although often cited in the literature (Hausteiner, Mergeay, Bornschein, Zilker, & Forstl, 2006), MCS is a more controversial designation (Henningsen et al., 2007). Medically diagnosed MCS is prevalent in 2.5% to 6.3% of the population (Caress, 2004; Kreutzer, Neutra, & Lashuay, 1999). All four FSS are thought to be more common in women as compared to men (Aggarwal, McBeth, Zakrzewska, Lunt, & Macfarlane, 2006; Johnson, 2008). FSS are a significant public health concern as they not only result in high costs to the health care system, but they can lead to the disruption of social roles, loss of productivity, and feelings of impotence and frustration on the part of health care providers (Barsky, Orav, & Bates, 2005; Centers for Disease Control and Prevention, 2006).

Childhood abuse has been implicated in a long list of serious health and social problems found to affect adults. These problems include physical illness, psychiatric disorders, substance abuse, and other high-risk health behaviors (Dube, Felitti, Dong, Giles, & Anda, 2003; Felitti et al., 1998; Heffernan et al., 2000; Kendall-Tackett, 2007; McCauley et al., 1997; Springer, Sheridan, Kuo, & Carnes, 2007).

In the Adverse Childhood Experiences (ACE) study, Anda and colleagues (2006) reported that one of the outcomes of adverse childhood experiences, which include multiple forms of abuse in combination with other measures of family dysfunction, was multiple somatic symptoms. Other research has supported the finding that adverse childhood experiences are associated with FSS, somatization, or medically unexplained symptoms (Drossman, Leserman, Toomey, & Hu, 1996; Johnson, 2008; Walker et al., 1999). Specifically, individuals with conditions such as CFS (Heim et al., 2006; Van Houdenhove et al., 2001), Fm (Boisset-Pioro, Esdaile, & Fitzcharles, 1995; Imbierowicz & Egle, 2003), IBS (Salmon, Skaife, & Rhodes, 2003), and MCS (Bell, Baldwin, Russek, Schwartz, & Hardin, 1998; Staudenmayer, Selner, & Selner, 1993) are significantly more likely to report that they were physically abused during their childhood than persons without those conditions.

In the ACE study, Anda and colleagues (2006) identified physical abuse as the most prevalent form of adverse childhood experience. Although there exists a burgeoning research literature on the varied adult health outcomes of childhood physical abuse (e.g., Felitti et al., 1998; Fuller-Thomson & Brennenstuhl, 2009; Goodwin & Stein, 2004; Kendall-Tackett, 2009) and some studies that have reported on a relationship between abuse and specific FSS (Salmon et al., 2003; Van Houdenhove et al., 2001), there has yet to be a particular focus on the relationship between physical abuse and FSS. However, to study the relationship between childhood physical abuse and FSS, a number of potentially confounding factors must be taken into account. Potentially confounding factors are variables shown to be associated with both childhood physical abuse and FSS that explain at least part of the abuse–FSS relationship. Based on a review of the literature, such factors could include other childhood adversities, adult health behaviors, adult socioeconomic status (SES) and stressors, and mental health.

OTHER CHILDHOOD ADVERSITIES

Childhood physical abuse often occurs in the context of other childhood adversities such as parental unemployment, parental mental illness and addictions, and parental divorce (Brown, Frederico, Hewitt, & Sheehan, 2000; Dong et al., 2004; Springer et al., 2007; Turner, Finkelhor, & Ormrod, 2007; Walsh, MacMillan, & Jamieson, 2002). In turn, childhood adversities such as parental addictions, parental conflict, poor relationships with parents, and emotional neglect have been associated with higher odds of several FSS (Imbierowicz & Egle, 2003; Van Houdenhove et al., 2001).

ADULT HEALTH BEHAVIORS

Felitti and colleagues (1998) associated negative health outcomes of adverse childhood experiences with adult health behaviors, including increased levels of smoking and alcohol use, physical inactivity, and obesity in adulthood (see also Anda et al., 2006; Dube et al., 2003). In FSS, obesity has been associated with Fm, although not etiologically (Yunus, Arslan, & Aldag, 2002), and research on exercise tolerance and CFS has been mixed (Bazelmans, Bleijenberg, Van Der Meer, & Folgering, 2001; Fischler et al., 1997). Other studies have reported inverse relationships between adverse adult health behaviors such as alcohol and drug use and CFS, Fm, and MCS, and between positive health behaviors such as exercise, and CFS (Harvey, Wadsworth, Wessely, & Hotopf, 2008; Hausteiner et al., 2006; Shaver, Wilbur, Robinson, Wang, & Buntin, 2006; Woolley, Allen, & Wessely, 2004).

ADULT SOCIOECONOMIC STATUS AND STRESSORS

Childhood physical abuse can lead to greater adversity in adulthood. Abused children are more likely to experience greater educational difficulties (Whiting, 2001), which potentially limits employment opportunities. As adults they are more likely to become divorced or separated (Colman & Widom, 2004). Childhood abuse has also been associated with higher perceived levels of stress (Anda et al., 2006). Similarly, adult socioeconomic indicators including lower educational level, lower household income, unemployment, and stressors such as being divorced or unmarried have been associated with higher odds of several FSS (Andrews et al., 2005; Walsh, Jamieson, MacMillan, & Boyle, 2007; Wolfe et al., 1995). Recent adverse life events and life stress are also highly associated with the onset of FSS including Fm, IBS, and CFS (Aggarwal et al., 2006; Hatcher & House, 2003).

MENTAL HEALTH

There are many established long-term negative mental health outcomes of abuse, including but not limited to depression and anxiety (Batten, Aslan, Maciejewski, & Mazure, 2004; Dube et al., 2003; McCauley et al., 1997; Springer et al., 2007). Moreover, meta-analytical research has demonstrated a consistent relationship between depression and anxiety, and CFS, Fm, and IBS (Henningsen, Zimmerman, & Sattel, 2003).

Increased understanding of the physical abuse–FSS relationship is important in light of the large numbers of individuals affected by FSS and the considerable personal, financial, and health impact of the syndromes. This study seeks to add to the literature on childhood physical abuse and adult illness by using a large, regionally representative sample of women to explore the relationship between childhood physical abuse and four FSS in adulthood (CFS, Fm, IBS, and MCS), while controlling for age and race in addition to the following four clusters of potential confounders: (a) Other childhood adversities, (b) adult health behaviors, (c) adult SES and stressors, and (d) mental health.

METHOD

Data Source and Sample

This study used the Public Use File data from the 2005 cycle (3.1) of the Canadian Community Health Survey (CCHS) conducted by Statistics Canada (2007). The CCHS is a cross-sectional survey that gathers data on the Canadian population related to health status, health care utilization, and health determinants. The survey is based on a multistage, stratified sampling design that is used to represent approximately 98% of the Canadian population ages 12 and over. A household-level response rate of 84.9% (n = 143,076) was achieved. Among the households sampled, 132,947 individuals responded, resulting in a person-level response rate of 92.9%. The combined national response rate based on household and individual levels was 78.9%. Data collection took place between January and December 2005. Each interview was about 45 min in length. Approximately half of the interviews were conducted in person, and the other half were conducted over the telephone. Both versions of interviewing were computer-assisted (Statistics Canada, 2006).

The analyses reported in this article are based on a subsample of 7,342 women from the Canadian provinces of Manitoba and Saskatchewan. Women were selected because they are more likely than men to suffer from FSS (Aggarwal et al., 2006; Barsky, Peekna, & Borus, 2001). This regional subsample was used for the analyses because the questions on childhood physical abuse were only asked in these two provinces. Questions on early childhood adversities were optional content to the survey that only certain regions chose to participate in. The combined provincial-level response rate for Manitoba and Saskatchewan was 83.3% and 84.1%, respectively (Statistics Canada, 2006). The CCHS did not include questions about other forms of childhood maltreatment, such as sexual abuse or emotional abuse and neglect.

Statistical Analyses

Six consecutive logistic regression analyses were conducted with CFS as the dependent variable and childhood physical abuse as the focal independent variable. The first and each subsequent model controlled for age and race. The second model adjusted for the first cluster of potentially confounding factors, namely other childhood adversities, including parental divorce, parental addictions, and parental unemployment. The third model adjusted for the second cluster, adult health behaviors, which included obesity, smoking, alcohol consumption, and recreational physical activity level. The fourth model adjusted for the third cluster, adult SES and stressors, including educational attainment, household income, daily self-reported stress level, and marital status. The fifth model adjusted for the fourth cluster, mental health problems, including a history of mood or anxiety disorders. The sixth and final model adjusted for age and race in addition to all four clusters of potentially confounding factors already identified. The six consecutive models were repeated for Fm, IBS, and then MCS.

For the purpose of clarity, we report in Figures 1 through 4 only the odds ratios associated with our key variable of interest: Self-reported childhood physical abuse. This allows us to focus the reader's attention on the way in which the association between childhood physical abuse and each FSS outcome varied with the inclusion of each cluster of potentially confounding factors.

FIGURE 1 Odds ratio and 95% confidence interval of irritable bowel syndrome for women reporting childhood physical abuse. All data are adjusted for age and race. Sample sizes vary from n = 7,274 in the first model to n = 7,067 in the fully adjusted model. Source: Representative, regional sample of the Canadian Community Health Survey (2005).

FIGURE 2 Odds ratio and 95% confidence interval of fibromyalgia for women reporting childhood physical abuse. All data are adjusted for age and race. Sample sizes vary from n = 7,276 in the first model to n = 7,070 in the fully adjusted model. Source: Representative, regional sample of the Canadian Community Health Survey (2005).

FIGURE 3 Odds ratio and 95% confidence interval of multiple chemical sensitivities for women reporting childhood physical abuse. All data are adjusted for age and race. Sample sizes vary from n = 7,272 in the first model to n = 7,068 in the fully adjusted model. Source: Representative, regional sample of the Canadian Community Health Survey (2005).

FIGURE 4 Odds ratio and 95% confidence interval of chronic fatigue syndrome for women reporting childhood physical abuse. All data are adjusted for age and race. Sample sizes vary from n = 7,276 in the first model to n = 7,072 in the fully adjusted model. Source: Representative, regional sample of the Canadian Community Health Survey (2005).

Due to missing data on one or more variables in the logistic regression models, sample sizes varied from the first model to the fully adjusted model. The final sample size of the fully adjusted model for each of the four FSS was as follows: CFS (n = 7,072), Fm (n = 7,070), IBS (n = 7,067), and MCS (n = 7,068). The maximum reduction in sample size from the original subsample (n = 7,342) to the final, fully adjusted model for any FSS was 3.7%. All analyses were weighted to adjust for the probability of selection and nonresponse.

Measures

Childhood physical abuse

Childhood abuse questions were in a section of the survey focusing on childhood and adult stressors. The following instructions preceded the section: “The next few questions ask about some things that may have happened to you while you were a child or a teenager, before you moved out of the house. Please tell me if any of these things have happened to you.” Women in the study were coded as having been physically abused as a child if they responded yes to the following question: “Were you ever physically abused by someone close to you?”

Functional somatic syndromes

Individuals were asked to identify if they had any “long-term conditions which are expected to last or have already lasted 6 months or more and that have been diagnosed by a health professional.” Using these instructions, individuals were asked these questions: “Do you have chronic fatigue syndrome?” “Do you have fibromyalgia?” “Do you have multiple chemical sensitivities,?” and “Do you suffer from a bowel disorder such as Crohn's disease, ulcerative colitis, irritable bowel syndrome, or bowel incontinence?” We identified those with IBS through their response to the subsequent question: “If yes, what kind of bowel disease do you have?” After every few questions, participants were reminded to report only those conditions that were diagnosed by a health professional.

Demographic characteristics

The demographic variables of age (18–29, then by decade to 80 and older) and race (White and visible minority) were also investigated. Unfortunately, information on the exact race and ethnicity of the visible minority respondents was not released in the CCHS Public Use Microdata File.

Other childhood adversities

Questions about other childhood adversities were found in the same section of the survey as the question about childhood abuse. Other childhood adversities investigated in this study included (a) parental divorce, as determined by the question, “Did your parents get a divorce?”; (b) parental unemployment, as determined by the question, “Did your father or mother not have a job for a long time when they wanted to be working?”; and (c) parental addiction, as determined by the question, “Did either of your parents drink or use drugs so often that it caused problems for the family?”

Adult health behaviors

Adult health behaviors including body mass index (BMI) category, smoking status, physical activity level, and alcohol use were investigated. BMI was calculated from self-reported height and weight and then classified into categories: (a) Normal weight (BMI < 25), (b) overweight (BMI = 25–29.9), or (c) obese (BMI = 30 or over). Smoking status was characterized into current, former and never smoker. Activity level was categorized into active, moderate, and inactive. Alcohol use was assessed by whether or not the respondent reported drinking five or more alcoholic beverages on one occasion, at least once in the last month.

Adult socioeconomic status and stressors

Adult SES included educational attainment and household income. Educational attainment was assessed by highest level of education: (a) Less than high school graduation, (b) high school graduation, and (c) postsecondary graduation. Household income was assessed according to the following income categories: (a) $0–$29,999, (b) $30,000–$49,999, or (c) $50,000 or more.

Two measures of adult stressors were examined: Self-reported daily stress and marital status. Daily stress was assessed by the question, “Thinking about the amount of stress in your life, would you say that most days are … ?” Response categories were dichotomized into not at all, not very, and a bit stressful versus quite a bit and very stressful. Marital status was used as a proxy for social support with the reasoning that unmarried women would experience less social support and in turn, greater life stress. Marital status was dichotomized into married or common law versus widow, separated, divorced, or single.

Mental health

Questions about mental health were found in the same section of the survey as the other long-lasting health conditions and were preceded by a reminder that interviewers were only interested in conditions diagnosed by a health professional. Depression was assessed by a self-reported response to the question: “Do you have a mood disorder such as depression, bipolar disorder, mania, or dysthymia?” Similarly, anxiety was assessed by self-report according to the question, “Do you have an anxiety disorder such as a phobia, obsessive-compulsive disorder, or a panic disorder?” Among the independent variables already listed, missing categories were constructed for the only two variables that were shown to have substantial missing data: BMI category and household income.

RESULTS

Using a representative regional subsample of the CCHS, the weighted prevalence of each of the four FSS investigated was found to be: (a) 1.3% (95% CI = 0.8, 1.8) ¹ for CFS, (b) 2.5% (95% CI = 1.7, 3.3) for Fm, (c) 4.2% (95% CI = 3.2, 5.2) for IBS, and (d) 2.7% (95% CI = 1.9, 3.5) for MSC. Approximately 10% of the women sampled reported being physically abused as a child. See Table 1 for a description of the sample. The final logistic regression model for each of the four FSS is found in Table 2.

TABLE 1 Description of a Regional Sample of Women from the Canadian Provinces of Manitoba and Saskatchewan Selected from the 2005 Canadian Community Health Survey (N = 7,342)

	N	%
Demographics
Age
18–29	1,313	17.9
30–39	1,041	14.2
40–49	1,087	14.8
50–59	1,194	16.3
60–69	1,001	13.6
70–79	960	13.1
80+ years	746	10.2
Ethnicity
White	6,391	87.8
Non-white	892	12.2
Childhood stressors
Parents unemployed for long time
No	6,587	90.8
Yes	666	9.2
Parents addicted to alcohol/drugs
No	6,144	84.0
Yes	1,172	16.0
Parents divorced
No	6,558	89.4
Yes	778	10.6
Adult health behaviors
Body mass index (BMI) category
BMI > 25	3,276	44.6
BMI 25–29.99	2,154	29.3
BMI 30 or greater	1,457	19.8
Missing	455	6.2
Smoker
Current	1,691	23.0
Former	2,919	39.8
Never	2,728	37.2
Physical activity level
Active	1,530	20.9
Moderate	1,815	24.7
Inactive	3,990	55.4
Binge drinking
No	6,602	90.2
Yes	719	9.8
Adult socioeconomic status and stressors
Education
< High school	1,863	25.6
High school/some postsecondary	1,937	26.6
Postsecondary	3,475	47.8
Amount of daily stress
None, not very, a bit stressful	5,849	80.2
Quite a bit, very stressful	1,448	19.8
Marital status
Single	3,456	47.2
Married	3,866	52.8
Household income
$0–$29,999	2,347	30.2
$30,000–$49,999	1,396	19.0
$50,000 or more	2,409	32.8
Missing	1,190	16.2
Mental health
Has mood disorder
No	6,786	92.5
Yes	552	7.5
Has anxiety disorder
No	6,922	94.4
Yes	413	5.6

TABLE 2 Logistic Regression Models of Self-Reported Childhood Physical Abuse and Functional Somatic Syndrome Diagnosis by a Health Professional

	Chronic fatigue^a		Fibromyalgia^b		Irritable bowel syndrome^c		Multiple chemical sensitivities^d
Independent variables	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
ExposurePhysical abuse as child
Yes	2.11	1.22, 3.65	1.65	1.08, 2.52	1.14	0.79, 1.65	2.82	1.90, 4.17
No	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
Control variables
Race
White	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
Visible minority	0.21	0.07, 0.64	0.25	0.11, 0.56	0.53	0.34, 0.83	0.57	0.32, 1.02
Age
18–29	0.92	0.29, 2.21	0.20	0.07, 0.61	2.44	0.89, 6.69	0.33	0.15, 0.73
30–39	1.38	0.43, 4.48	0.60	0.23, 1.57	3.23	1.17, 8.86	0.26	0.11, 0.60
40–49	1.64	0.52, 5.20	1.39	0.57, 3.40	3.10	1.13, 8.48	0.70	0.34, 1.43
50–59	1.40	0.44, 4.50	1.43	0.59, 3.51	3.00	1.10, 8.22	0.85	0.42, 1.70
60–69	1.96	0.61, 6.36	1.95	0.81, 4.72	2.35	0.84, 6.63	1.37	0.70, 2.68
70–79	1.2	0.36, 4.07	1.12	0.43, 2.89	2.16	0.74, 6.27	1.07	0.54, 2.13
80+	01.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
Covariates
Childhood stressors
Parental divorce
Yes	2.13	1.21, 3.76	1.11	0.68, 1.80	0.86	0.58, 1.28	0.49	0.26, 0.92
No	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
Parental addictions
Yes	0.61	0.33, 1.12	1.22	0.83, 1.80	0.86	0.62, 1.20	1.38	0.94, 2.04
No	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
Parental unemployment	2.78	1.63, 4.76	1.49	0.94, 2.38	1.67	1.18, 2.38	0.84	0.49, 1.43
Yes	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
No
Adult health behaviors
Smoking
Current	0.60	0.32, 1.15	1.79	1.16, 2.78	1.07	0.77, 1.48	0.66	0.43, 1.02
Former	1.33	0.81, 2.19	1.49	1.02, 2.18	1.03	0.78, 1.36	0.84	0.60, 1.16
Never	1.00	Referent	1.00	Referent	Referent	Referent	1.00	Referent
BMI category
BMI > 25	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
BMI 25–29.99	0.73	0.42, 1.26	0.93	0.64, 1.35	0.92	0.69, 1.23	1.27	0.90, 1.79
BMI 30 or greater	1.35	0.79, 2.34	1.57	1.07, 2.30	1.03	0.74, 1.42	1.12	0.75, 1.67
Missing	0.29	0.06, 1.36	0.99	0.43, 2.27	0.62	0.31, 1.23	1.24	0.61, 2.52
Physical activity level
Active	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
Moderate	0.58	0.30, 1.11	0.88	0.54, 1.45	1.26	0.90, 1.75	1.01	0.60, 1.71
Inactive	0.82	0.48, 1.40	1.21	0.79, 1.86	0.76	0.55, 1.05	1.73	1.11, 2.71
Binge drinking
Never or < 1 time/month	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
1 or more times/month	1.13	0.55, 2.33	1.10	0.62, 1.96	0.77	0.50, 1.17	0.61	0.30, 1.26
Adult SES and stress
Education level
< High school (HS)	1.01	0.52, 1.96	0.73	0.46, 1.16	0.62	0.40, 0.94	0.95	0.63, 1.45
HS/some postsecondary	1.81	1.13, 2.91	1.03	0.73, 1.46	0.73	0.55, 0.97	1.14	0.81, 1.61
Postsecondary	1.00	Referent	1.00	Referent	1.00	1.33, 2.23	1.00	Referent
Income
Missing	3.06	1.60, 5.85	0.79	0.45, 1.39	0.68	0.43, 1.06	0.60	0.35, 1.05
$0–$29,999	2.21	1.17, 4.18	1.38	0.87, 2.19	0.96	0.65, 1.41	0.97	0.62, 1.52
$30,000–$49,999	2.09	1.13, 3.89	1.33	0.89, 1.99	1.57	1.17, 2.11	1.19	0.80, 1.76
$50,000 or more	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
Self-reported stress level
Not at all/a bit	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
Quite a bit/extremely	3.08	1.96, 4.82	1.30	0.93, 1.82	1.73	1.33, 2.23	1.40	1.01, 1.95
Marital status
Married	1.00	Referent	1.00	Referent	1.0	Referent	1.00	Referent
Unmarried	1.37	0.85, 2.22	0.70	0.48, 1.02	01.01	0.76, 1.34	0.96	0.68, 1.36
Adult mental health
Has a mood disorder
Yes	1.24	0.67, 2.30	2.00	1.27, 3.14	1.69	1.18, 2.42	1.82	1.16, 2.85
No	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
Has an anxiety disorder
Yes	3.92	2.15, 7.12	1.69	1.03, 2.77	2.70	1.87, 3.89	2.28	1.43, 3.64
No	1.00	Referent	1.00	Referent	1.00	Referent	1.00	Referent
Note: OR = odds ratio; CI = confidence interval; BMI = body mass index.
^a n = 7,072. ^b n = 7,070. ^c n = 7,067. ^d n = 7,068.

To focus the reader's attention on the abuse–FSS relationship, we report in Figures 1 through 4 only the odds ratios associated with our key variable of interest: Self-reported physical abuse. Childhood physical abuse was associated with higher odds of all four FSS when controlling for age and race. The highest odds associated with childhood physical abuse were for CFS (OR = 4.17, 95% CI = 2.64, 6.59), followed by MCS (OR = 3.15, 95% CI = 2.21, 4.49), and Fm (OR = 2.44, 95% CI = 1.67, 3.58). When adjustments were made for the four clusters of risk factors, childhood physical abuse remained a significant factor for CFS (OR = 2.11, 95% CI = 1.22, 3.65), MCS (OR = 2.82, 95% CI = 1.90, 4.17), and Fm (OR = 1.65, 95% CI = 1.08, 2.52). Adult SES and life stressors, followed by mental health and other childhood adversities, each resulted in a decrease in the physical abuse–CFS relationship. The abuse–Fm relationship decreased the most when mental health or other childhood adversities were included. The physical abuse–MCS relationship declined the most when mental health or adult stressors were included in the analysis.

The age- and race-adjusted odds of IBS among those with childhood physical abuse were significantly elevated (OR = 1.52, 95% CI = 1.09, 2.12), but it was considerably lower than the other FSS. Including either mental health or adult SES and stressors in the equation reduced the abuse–IBS relationship to nonsignificance. Adjusting for adult health behaviors had virtually no impact on the abuse–FSS relationship for any of the four types of FSS investigated.

DISCUSSION

This study examined the relationship between FSS and childhood physical abuse controlling for other childhood adversities, adult health behaviors, adult SES and life stressors, and mental health. Much of the FSS literature to date has focused on the importance of depression and anxiety as potential mediators. We found that adjusting for anxiety and mood disorders reduced the magnitude of the abuse–FSS relationship for Fm, IBS, and MCS more than any of the other clusters of potentially confounding factors analyzed. Clearly, depression and anxiety are important variables to control for when studying FSS among women who have been abused. However, for three of the four syndromes examined, a significant and robust abuse–FSS relationship remained even after adjusting for such well-established risk factors as a history of mental illness. In fact, in the fully adjusted model (as shown in Table 2), the magnitude of the association between childhood physical abuse and FSS was comparable to the association between a history of mood disorders and FSS for CFS (OR for abuse = 2.11; OR for mood disorder = 1.24), Fm (OR for abuse = 1.65; OR for mood disorder = 2.00), and MCS (OR for abuse = 2.82; OR for mood disorder = 1.82).

Adjustment for adult health behaviors had little impact on the abuse–FSS relationship. This is not surprising because, to our knowledge, smoking, alcohol use, other substance use, physical inactivity, and obesity have not been clearly associated with the onset of FSS.

Other childhood adversities played a more substantial role in reducing the abuse–CFS and the abuse–Fm relationship than it did in the other two syndromes. This is expected given the greater frequency of early stressors such as parental alcoholism in Fm (Imbierowicz & Egle, 2003) and parental conflict, poor relationship with parents, and emotional abuse and neglect reported in CFS and Fm (Imbierowicz & Egle; Van Houdenhove et al., 2001).

Including adult SES and stressors in the analysis resulted in a large decline in the abuse–CFS relationship. This finding is consistent with the strong evidence of the role that recent adverse life events and life stress play in the onset of FSS (Aggarwal et al., 2006; Hatcher & House, 2003; Salit, 1997).

Different Relationships between Physical Abuse and Each FSS

In this analysis, the four FFS differ in their relationship to childhood physical abuse. For MCS, controlling for all four clusters of potentially confounding factors combined had very little impact on the relationship between childhood physical abuse and MCS. Once we considered the impact of other childhood adversities and mental health on Fm, the relationship between childhood physical abuse and Fm declined but remained significant. In CFS, adjusting for other childhood adversities, adult SES and life stressors, and mental health all resulted in a substantial decline in the child physical abuse–CFS relationship. However, as with MCS and fibromyalgia, abuse remained a significant factor. In contrast, the inclusion of mental health reduced the relationship between childhood physical abuse and IBS to nonsignificance. This suggests that the childhood physical abuse–IBS relationship might primarily be due to the association between childhood physical abuse and higher rates of anxiety and mood disorders. Anxiety and depression have been significantly associated with IBS in this and in other research (Creed et al., 2005; Drossman et al., 2000; Lydiard, 2001; Sayuk, Elwing, Lustman, & Clouse, 2007). The relationship between abuse and IBS also declined when controlling for adult SES and life stressors. In our study and in other studies, life stress (Aggarwal et al., 2006; Locke, Weaver, Melton, & Talley, 2004), SES (Lerebours et al., 2007), and marital status (Andrews et al., 2005) were significantly associated with IBS. Our study and Henningsen et al.'s (2003) meta-analysis indicated that FSS are related to but not fully dependent on depression and anxiety. Further exploration is needed to understand the differences in the relationship between childhood physical abuse and each of the four FSS.

Potential Mechanisms Linking Physical Abuse to FSS

To our knowledge, this study of the physical abuse–FSS relationship controlled for a wider range of potentially confounding factors than previous research using representative community-based samples. Despite controlling for 15 variables, the abuse–FSS relationship remained large and significant for CFS, Fm, and MCS. It appears that there must be additional factors that influence the abuse–FSS relationship.

Several researchers have postulated mechanisms that mediate or moderate the impact of childhood physical abuse on adult health. Using attachment and personal control theory, Shaw and Krause (2002) proposed that childhood physical abuse can impair the development of a sense of control and the ability to form supportive social relationships, both of which are “intervening mechanisms … also strongly related to physical health” (p. 474). Anda and colleagues (2006) and Dong, Dube, Felitti, Giles, and Anda (2003) documented risk behaviors ensuing from a disordered social environment that can negatively affect the developing brain and act as causal pathways from adverse childhood experiences to illness in adults.

However, other researchers have suggested that the increased incidence of adult health conditions including FSS among those who experienced childhood abuse could be partially a result of the early biological embedding (Hertzman, 1999) of health risks (Anda et al., 2006; Weissbecker, Floyd, Dedert, Salmon, & Sephtn, 2006). Biological embedding refers to the process by which early life experiences such as poverty or childhood physical abuse “establish functional and structural changes in the physiology, neurobiology, and gene expression of the individual,” which subsequently increases the person's vulnerability to a wide range of health problems (Kuh & Ben-Shlomo, 2004, p. 421). In addition, prolonged exposure to traumatic events in childhood might result in the heightened sensitivity of the hypothalamus–pituitary–adrenal (HPA) axis to stress throughout the life span (Crofford, 2007; Hertzman). The HPA axis plays a primary role in the production of cortisol in response to stressful events (Hertzman). Clinical studies have demonstrated that ongoing or significant stress during development, including childhood abuse, can result in the neuroendocrine dysregulation of the adult stress response (Heim & Nemeroff, 2002; Heim et al., 2002; Stein, Yehuda, Koverola, & Hanna, 1997). Among adults with Fm, childhood abuse has been associated with significant neuroendocrine dysregulation (Weissbecker et al., 2006).

Limitations

There are several limitations to this research. Due to the cross-sectional design of the CCHS, we cannot draw any conclusions on causality. Longitudinal prospective studies are necessary to clarify the abuse–FSS relationship and the role of mediating factors. Furthermore, childhood physical abuse was identified through respondent recall rather than through objective measures. In particular, the question used to ascertain whether abuse had occurred was not well defined for the participants, allowing each woman to determine subjectively which, if any, of her experiences constituted abuse. Nonetheless, due to the frequency of FSS seen in primary care, the perception of being physically abused as a child has clinical importance when it is linked with these illness outcomes. Health professionals rarely have access to information on early childhood abuse other than self-report. Therefore, we believe that the relationship we have found between self-reported abuse and FSS might be of value to clinicians.

It is possible that adults with FSS or other risk factors might be more or less likely to report childhood physical abuse (Felitti et al., 1998; Springer et al., 2007). Previous longitudinal research has shown that retrospective reports of childhood abuse are more likely to underestimate actual occurrences (Fergusson, Horwood, & Woodward, 2000). The prevalence of self-reported physical abuse among Canadian women of 9.6% found in this study is much lower than a previous Canadian community study of 21.1% of females, although severe physical abuse was reported by a similar proportion (9.2%; MacMillan et al., 1997). The percentage in this study is also lower than in American studies such as that of Afifi, Brownridge, Cox, and Sareen (2006), who found that 16.5% of their national sample reported affirmatively to a much more specific list of serious abuses such as being “kicked, bit, hit with a fist, hit or tried to hit with something, beat up, choked, burned, or scalded” (p. 1096). This suggests that given the broader scope of the question about childhood physical abuse in this study, there is a likelihood of underreporting.

Classification of each type of FSS was based on the respondents' report of a diagnosis by a health professional, which could result in misreporting. For example, less than 20% of Americans with CFS have been diagnosed (Centers for Disease Control and Prevention, 2006). Future research would benefit from the use of objective medical criteria to determine FSS. Finally, several important factors were not available in the CCHS data set. Because adverse childhood experiences tend to cooccur (Dong et al., 2004), it would have been helpful to control for childhood sexual abuse and other known risk factors for FSS. Moreover, our conceptualization of some groups of factors, such as adult stressors, was constrained by the variables available in the CCHS. For example, data on stressors such as lack of social support were not available, thus requiring us to use a proxy (i.e., marital status). However, it is uncertain how well marital status represents overall levels of social support.

Implications for Clinicians

Reattribution and stepped care approaches with patients who have medically unexplained symptoms offer encouraging trends in primary care management of FSS (Fink & Rosendal, 2008). According to Morriss and colleagues (2006), reattribution is a structured intervention, designed to provide a simple explanation of the mechanism of a patient's medically unexplained symptoms, which can be delivered during routine consultations. It has four stages. In Stage 1, feeling understood, the primary care clinician uses empathic listening skills to take a history of physical, emotional, and psychosocial factors of the presenting symptoms, including the patient's beliefs and perceptions (e.g., causality, when it's worse, what helps), and completes a brief, focused physical. In Stage 2, broadening the agenda, the care provider gives feedback and implications of the findings, acknowledges the patient's distress, provides realistic reassurance, and reflects back emotional and psychosocial cues that the patient has given during the assessment (Guthrie, 2008). In Stage 3, making the link, the care provider works with the patient's cues to provide an empowering explanation (i.e., one that links symptoms to cues and provides a physiological mechanism that removes any sense of blame from the patient; Salmon, Peters, & Stanley, 1999). In Stage 4, negotiating further treatment, the care provider and patient join to create a treatment plan and to decide whether follow-up is needed.

Guidelines for the range of treatments are suggested in the concept of stepped care. These guidelines move in stepped fashion from mild and uncomplicated symptoms to severe and complex issues. Persons presenting with mild and moderate uncomplicated symptoms will likely respond to empowering explanations, reassurance, and appropriate behavioral advice, or pharmacotherapy. Those with moderate or severe complex symptoms are more likely to require multidisciplinary consultation or referral for intervention (Fink & Rosendal, 2008; Henningsen et al., 2007). Examples of treatments that might be helpful are cognitive-behavioral therapy and other types of psychosocial therapies, physiotherapy and rehabilitation, relaxation training and stress management, nursing, nutritional counseling, and clinical pharmacology. For patients with FSS, we recommend that clinicians who are using reattribution and stepped care methods assess for childhood physical abuse as well as other adverse childhood experiences, as these issues might not readily surface in a primary care visit unless they are raised by the care provider.

Conclusions

This study provides strong evidence of a relationship between childhood physical abuse and FSS using a representative community-based sample. Childhood physical abuse was significantly associated with higher odds of CFS (OR = 2.11), Fm (OR = 1.65), and MCS (OR = 2.82), even when controlling for other childhood adversities, adult health behaviors, adult SES and stressors, and mental health. When controlling for these factors, childhood physical abuse was no longer significantly associated with IBS (OR = 1.14). Mood and anxiety disorders might be a potential mediating pathway through which physical abuse in childhood is associated with IBS. Future community-based research should attempt to use a more objective assessment of FSS and control for other forms of abuse, especially sexual abuse. Finally, longitudinal research is needed to determine the mechanisms that connect childhood physical abuse to FSS in adulthood.

Notes

¹This percentage is associated with a high sampling variability.