| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From UMDNJ–Robert Wood Johnson Medical School, Piscataway, New Jersey.
Address reprint requests to: Dr. Alejandro Interian, Department of Psychiatry, UMDNJ—Robert Wood Johnson Medical School, 675 Hoes Lane D306, Piscataway, NJ 08854-5635.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: Interview and questionnaire data were obtained from 120 patients with somatization who participated in a study assessing the efficacy of cognitive-behavioral therapy. Measures elicited information on psychiatric diagnoses, anxiety and depressive symptom levels, somatic symptoms, and physical functioning. Statistical analyses examined the relationship between PNS and the diagnosis of somatization disorder, physical and psychiatric symptom severity, and psychiatric comorbidity.
RESULTS: Roughly half of the sample had a history of four or more PNS. Results showed that having four or more PNS was not predictive of somatization disorder. However, having four or more PNS was found to be significantly correlated with the severity of anxiety, depression, somatic complaints, and physical dysfunction. These associations were identified while controlling for the symptom count of nonpseudoneurological symptoms, the presence of somatization disorder, and the presence of chronic painful physical conditions. In addition, having four or more PNS was significantly associated with a higher likelihood of receiving a diagnosis of major depression, dysthymia, panic disorder, and generalized anxiety disorder.
CONCLUSIONS: A history of four or more PNS is common among somatizing patients in primary care and associated with a more severe clinical presentation, even after controlling for other factors known to be associated with severity. Four or more PNS may identify a distinct subgroup of somatization and serve as a clinical indicator for identifying psychiatric disorders in primary care. Future studies should explore the assessment of PNS using briefer measures. Furthermore, PNS should be evaluated with samples more representative of US primary care populations, as well as samples that include adequate representation from other ethnic backgrounds (eg, African-American, Asian, etc.).
Key Words: pseudoneurological symptoms, • primary care, • somatization, • somatic symptom clusters, • psychiatric diagnosis.
Abbreviations: CGI = Clinical Global Impression;; CIDI = Composite International Diagnostic Interview;; HAM-A = Hamilton Anxiety Scale;; HAM-D = Hamilton Depression Scale;; MOS SF-10 = RAND MOS Short-Form Health Survey;; PHQ-13 = Patient Health Questionnaire somatic module;; PNS = pseudoneurological symptoms;; PRIME-MD = Primary Care Evaluation of Mental Disorders;; SD = somatization disorder.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
From a healthcare perspective, somatization is a frustrating and costly disorder to treat and manage, as these patients are avid utilizers of healthcare and tend to be chronically dissatisfied with their care (12–15).
In DSM-IV (16), suitable diagnoses for patients with multiple unexplained physical symptoms are listed under the Somatoform Disorders category and include Somatization Disorder and Undifferentiated Somatoform Disorder. Criteria for "full" Somatization Disorder require a history of at least 8 unexplained physical symptoms that originate from multiple organ systems (eg, musculoskeletal, gastrointestinal). Somatization Disorder (SD) identifies patients with severe forms of somatization who are quite uncommon in community and clinical settings (13). However, Abridged Somatization, a less stringent dimension of somatization defined by Escobar et al. (13), requires 4 unexplained symptoms for males and 6 for females and has been found to be more common in community samples. At the same time, abridged criteria identify individuals with clinical characteristics similar to patients with SD, such as high rates of healthcare utilization and comorbid psychiatric disorders (5).
"Full" Somatization Disorder and Abridged Somatization capture different degrees of severity that are associated with variable levels of psychiatric morbidity and physical dysfunction. These two definitions constitute cross-sectional cut-offs (symptoms counts) at different points on a continuum of somatic symptoms (2, 5). In addition to symptom counts, other dimensions have been shown to highlight differences between somatizers. One such dimension is the number of organ systems (eg, pseudoneurological, gastrointestinal, cardiorespiratory, etc.) from which the unexplained symptoms originate. In one study, the greater the number of organ systems affected, the more physical disability was found (5). Escobar et al. also proposed that the type of organ system be used to classify somatizing patients based on symptom clusters.
Somatic Symptom Clusters
One study examined the potential use of symptom clusters for identifying subgroups of patients with somatization (17). The study utilized numerical clustering methods to examine patients from a general primary care sample. A particular variant of cluster analysis, Hierarchical Classes Analysis (HICLAS), was used to identify groupings of patients based on the type of organ system implicated in their unexplained symptoms. Their results showed that the type of organ system could be used to form distinct clusters of patients, including pseudoneurological, gastrointestinal, musculoskeletal, cardiorespiratory, and other symptoms.
Pseudoneurological Symptoms
Of the somatic symptoms listed in DSM-IV, pseudoneurological symptoms (PNS) have a long and distinguished tradition in psychopathology. Indeed, the original construct of Hysteria/Briquet’s Syndrome was made up exclusively of PNS. Thus, while newer classifications have added many other symptoms, we believe that patients with PNS represent a unique group.
In this regard, findings reported by Gara et al. (17) suggested that the pseudoneurological cluster demonstrated a number of unique characteristics. While most patients in the study had symptoms from 1 or 2 organ cluster(s)—rarely from 3 clusters—patients in the pseudoneurological cluster were unique because they presented not only with PNS, but also with symptoms from virtually all other organ systems. This pattern of multiple symptom clusters for patients with PNS is supported by a prior study that utilized a community sample and a similar cluster analytic technique (18). The results of Gara et al.’s cluster analysis also showed that patients with PNS were distinct in 2 additional ways. First, patients with 4 or more PNS also met criteria for both Abridged and "full" Somatization Disorder. Second, patients with 4 or more PNS represented a more severe group of patients, as indicated by increased psychiatric comorbidity, a greater number of anxious and depressive symptoms, and greater physical disability. In other words and as noted by the authors, PNS identified a group of patients who represented the more severe end of the somatization and psychopathology continuum.
From the perspective of primary care practitioners, the results of Gara et al. suggest that PNS can provide important clinical data for the detection of psychopathology. Since our study focused on somatization in primary care, we sought to evaluate the utility of Gara et al.’s proposed criterion of 4 or more PNS in a new cohort. Our focus was on 4 or more PNS because previous results showed this criterion to be the optimal predictor.
Using data in the current sample, the PNS criterion was first examined for its ability to concurrently predict Somatization Disorder. Other analyses on the current data set examined whether patients with 4 or more PNS evidenced greater severity in general psychopathology by measuring anxiety and depressive symptoms and syndromes, as well as functional status. The overall intent of the analysis was to evaluate the suitability of PNS as a construct for a somatization subtype and a clinical tool for detecting psychiatric disorders and somatization severity.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Measures
Patient Health Questionnaire
First, participants completed the entire Patient Health Questionnaire (PHQ) (19). In the current analysis, however, only the 13 somatic symptoms (PHQ-13) from the somatic symptom module were utilized. The PHQ-13 served as an index of somatic complaints during the month before the evaluation and demonstrated acceptable internal consistency (
= 0.76) with the current sample.
Primary Care Evaluation of Mental Disorders
The presence of depressive, anxious, and eating disorders during base-line was evaluated using the Primary Care Evaluation of Mental Disorders (PRIME-MD), which is a structured clinical interview (20). The PRIME-MD also has a patient questionnaire that was used as an initial screen for multiple medically unexplained symptoms.
Composite International Diagnostic Interview
The participants’ history of 41 unexplained physical symptoms was evaluated using the somatization module of the Composite International Diagnostic Interview (CIDI) (21). The CIDI is administered in interview format and contains a structured set of probes that directs the interviewer in determining whether those symptoms were significantly distressing and medically unexplained. Referring physicians and/or nurses were consulted regarding symptoms for which the presence of a medical explanation was unclear.
Clinical Global Impression
Participants’ somatic presentations were also rated using the Clinical Global Impression (CGI). This is a 1-item index that allowed the clinicians to rate the global level of somatization severity on a 7-point scale. The CGI utilizes anchors pertaining to total symptom count and degree of impairment.
Hamilton Depression Scale and Hamilton Anxiety Scale
The Hamilton Depression Scale (HAM-D) (22) and Hamilton Anxiety Scale (HAM-A) (23) were also administered in interview format. The HAM-D contained 17 items, while the HAM-A contained 14 items. Clinicians were trained on administering the HAM-D and HAM-A. During the first year of data collection, 6 interviews (3 Spanish; 3 English) were randomly selected to be re-rated by all clinicians using audiotape. Each interview was re-rated by 1 clinician who provided HAM-D and HAM-A ratings and pairs of raters were alternated randomly. Kappa statistics were used to measure the interrater agreement of these pairs (HAM-D 
= 0.81; HAM-A = 0.68).
RAND MOS Short-Form Health Survey
For functional status data, participants completed the RAND MOS Short-Form Health Survey (MOS SF-12) (24), which is a self-report questionnaire that assesses overall perception of their health and physical functioning. The analysis utilized the sum of the 10 behavioral items of the measure (MOS SF-10), which queried participants about their ability to perform certain physical tasks (eg, climb stairs, walk more than 1 mile, etc.). The MOS SF-10 demonstrated acceptable internal consistency with the current sample ( = 0.69).
Design
As noted, data were obtained from participants who were enrolled in a clinical trial. Thus, the current study was a secondary analysis that used base-line data from all participants enrolled at that point in time. The sample was one of convenience, where recruitment was dependent on affiliations with the outpatient clinics and the referral patterns within those clinics. Also, one of the clinics served a largely non-English speaking, Hispanic population.
The presence of PNS, other medically unexplained symptoms, and Abridged and "Full" Somatization disorders was assessed with the somatization module of the CIDI. The 41 somatization items included 15 individual PNS (eg, blindness, paralysis, balance difficulty). Using the PNS criterion proposed by Gara et al. (17), patients with at least 4 PNS were identified as part of the PNS group. All others were identified as part of the non-PNS group.
Procedure
Participants were referred by primary care physicians and nurses if they presented with a history of at least 3 unexplained physical symptoms, as indicated by participants’ responses on the PRIME-MD patient questionnaire. Participants were subsequently prescreened by telephone to ensure they met basic criteria for the clinical trial. Inclusion criteria required somatic presentations that fulfilled Escobar’s definition of Abridged Somatization, as later assessed by somatization items of the CIDI. Participants with psychotic symptoms, active suicidal ideations, active substance abuse/dependence, acute medical illness, concurrent legal issues, concurrent psychosocial treatment, or pharmacological treatment initiated less than 3 months before enrollment were excluded to ensure the selection of "primary" somatizers. After prescreening, clinical interviews were performed by 4 doctoral level psychologists and 1 psychology doctoral candidate. These clinicians collected basic demographic data and administered the PRIME-MD, CIDI somatization module, HAM-D, and HAM-A. Participants subsequently completed the MOS SF-12 and PHQ. After the interview, clinicians rated the severity of somatization on the CGI. The PRIME-MD was included as a base-line, clinician-administered measure of diagnostic status, while the PHQ was used as a self-report outcome measure.
Of a total of 328 participants referred to the study, 120 were enrolled and 208 were excluded for the following reasons: 1) insufficient somatic symptoms (N= 74); 2) participant declined (N= 41); 3) scheduling difficulties (N= 28); 4) acute medical illness (N= 19); 5) concurrent treatment with psychotropic medications or psychotherapy (N= 18); 6) significant psychiatric disturbance (N= 11); and 7) significant legal issues (N= 7). The remainder of participants was not enrolled for reasons that included leaving the US and being pregnant.
Data Analyses
To assess whether PNS concurrently predicted the presence of Somatization Disorder, a cross-tabulation and Fishers Exact Test were conducted.
Data analyses also focused on the differences between the PNS and non-PNS groups with respect to the HAM-D, HAM-A, CGI, MOS SF-10, and PHQ-13. Analyses of Covariance (ANCOVA) were conducted to assess group severity differences on these dependent measures, while controlling for other variables associated with higher severity. Specifically, the number of nonpseudoneurological symptoms (eg, gastrointestinal, pain, cardiorespiratory) was controlled for in the ANOVA model to ensure that symptom count was not a confounding variable. Furthermore, Somatization Disorder, also shown to be associated greater severity (5, 25), was entered as an effect in the ANCOVA model. In essence, the ANCOVA model was used to examine whether effects related to the PNS construct were independent from other severity factors.
Finally, the likelihood of receiving comorbid psychiatric diagnoses on the PRIME-MD was evaluated using a cross-tabulation procedure and Fishers Exact Test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
For each participant, we calculated the percentage of medically unexplained symptoms that were pseudoneurological. For participants in the sample, the percentage of symptoms that were pseudoneurological ranged from 0 to 66.67% with a mean of 28.6%. Using the criterion of 4 or more PNS, 55 (46.22%) participants were classified into the PNS group. Table 1 compares demographic features of the PNS and non-PNS groups. No significant differences between the groups were found on any of the demographic variables. Table 2 summarizes the frequencies of PNS within both groups and presents p values from Fisher’s Exact Tests. For the PNS group, the most frequently presented PNS were balance difficulty, fainting spells, and numbness and tingling sensations, while the least frequent PNS were double vision, amnesia, and deafness. With the exception of temporary deafness and loss of voice, all PNS symptoms were significantly more frequent among participants in the PNS group.
|
|
ANCOVAs did however show that having 4 or more PNS was significantly related to more severe scores on the HAM-D, F (1, 115) = 10.63, p< .002); HAM-A, F (1, 115) = 8.72, p< .01); CGI, F (1, 115)= 14.01, p< 0.001); MOS SF-10, F (1, 110) = 5.08, p< .05); and PHQ-13, F (1, 115)= 7.45, p< .01). These ANCOVAs showed that membership/nonmembership in the PNS group is significantly related to each of the dependent measures, even after controlling for the number of nonpseudoneurological symptoms and the presence/absence of Somatization Disorder. The severity score differences between the PNS and non-PNS groups are summarized in Table 3. Adjusted p values represent statistical significance after controlling for the number of nonpseudoneurological symptoms and the presence of Somatization Disorder, while unadjusted p values refers to analyses without covariates. The differences between adjusted and nonadjusted p values are not large, which suggests that effects of PNS are independent from those of the other variables.
|
Finally, Figure 1 presents the percentage of participants within each group that were assigned a DSM-IV diagnosis based on the PRIME-MD. Overall, participants in the PNS group were more likely than patients not in the PNS group to present with any of the DSM-IV diagnoses included in the PRIME-MD. Fisher’s Exact Tests showed these co-morbidity patterns to be significant with respect to 4 Axis I disorders. Specifically, patients meeting the PNS criterion were significantly more likely than patients without PNS to be diagnosed with current major depression (p= .002), dysthymia (p= .009), panic disorder (p= .003), and generalized anxiety disorder (p= .002).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
While the results reported by Gara et al. (17) showed that PNS predicted DSM-IV Somatization Disorder (SD), this was not replicated in the current study. In fact, current results showed that the PNS criterion performed no better than chance in predicting SD. This negative result seems explained by the fact that the current sample comprises only somatizing patients, whereas Gara et al. analyzed a general primary care sample that included a large percentage of nonsomatizing participants. This finding provides some support for the construct validity of PNS, as it suggests that patients with four or more PNS form a somewhat independent subgroup.
Other analyses provided further support for the PNS construct and its potential to identify a subgroup of somatizing patients with rather distinct characteristics. Patients included in this subgroup have higher indexes of severity on a number of clinical and functional dimensions. PNS patients were also were also rated as having more severe presentations of somatization. These relationships held even after controlling for other variables known to be related to severity (5, 25, 26), such as the number of nonpseudoneurological symptoms, the presence/absence of Somatization Disorder, and the presence/absence of Chronic Painful Physical Conditions. Gara et al. (17) had demonstrated a significant relationship between PNS and psychiatric comorbidity, where PNS patients showed increased risks for major depression, anxiety disorder, dysthymic disorder, and melancholia. Our results also support those previous findings, specifically for major depression, dysthymia, panic disorder, and generalized anxiety disorder.
An important new feature of the current analyses is that we controlled for the effects of variables related to severity (ie, nonpseudoneurological symptoms, the presence of Somatization Disorder, and the presence of Chronic Painful Physical Conditions) and demonstrated that PNS predict key outcomes independent from these indicators. This highlights the relevance of the PNS subtype for clinical studies.
Taken together, the results of this study and those previously reported have important implications for the assessment of mental disorders in primary care. There has been abundant research documenting that mental disorders are traditionally undetected in primary care settings (20, 27). Some evidence also suggests that these low detection rates are further adversely affected by the presence of somatization and also by certain ethnic backgrounds, such as Hispanic (28, 29). In light of this, the PNS construct should be developed into a tool for primary care physicians, as there is promise for detecting mental disorders particularly among minority populations, such as Hispanic patients. Given the high levels of psychiatric comorbidity among patients with PNS, the presence of these symptoms can serve as a useful clinical indicator that prompts further questioning regarding psychiatric symptoms. Therefore, it is recommended that studies explore efficient methods for assessing these symptoms via brief measures. Currently, only subsets of PNS are included in the various brief somatic symptoms measures (eg, PHQ (19), Somatoform Disorders Schedule (30), & Hopkins Symptom Checklist (31)) and the PHQ and SCL-90 do not differentiate between medically explained and unexplained. PNS can be combined with other detection tools that have been described to serve a similar purpose. For example, recent stress, 5 or more physical symptoms, or poor health have been reported to be useful clinical indicators of psychiatric comorbidity (32).
In understanding the differences among somatizing patients, PNS appear to provide a useful template for recognizing intragroup differences. With psychotherapeutic and psychopharmacologic interventions recently having been studied with this group (33–37), PNS could serve as a useful variable for evaluating differential treatment response.
Limitations
The current study has several limitations. First, the sample is one of convenience with a preponderance of Latino Spanish-speaking female patients. While this provided an important opportunity to study this population, the findings may not be representative of other primary care samples. Moreover, a larger number of non-Latinos would have facilitated the determination of any cross-ethnic differences vis-a-vis somatization and the role of pseudoneurological symptoms. In addition, unexplained physical symptoms were assessed in a structured interview (CIDI), which is subject to a variety of interviewer and self-report biases. At the same time, however, physicians were consulted in instances where determination of a symptom’s medical explanation was difficult, which was a strength of this study. Finally, psychiatric diagnoses were rendered using the PRIME-MD, which has limitations relative to more traditional structured assessments such as the SCID-I (38). However, the PRIME-MD is more practical in the busy primary care research setting. Replication of the current results samples more representative of US primary care populations, with samples represented by other ethnic backgrounds (eg, African-American, Asian, etc.), and with other assessment methods is in order.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication May 27, 2003.
Revision received October 3, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. Noyes Jr., S. P. Stuart, and D. B. Watson A Reconceptualization of the Somatoform Disorders Psychosomatics, February 1, 2008; 49(1): 14 - 22. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. J. Kirmayer and N. Sartorius Cultural Models and Somatic Syndromes Psychosom Med, November 1, 2007; 69(9): 832 - 840. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Rief and G. Rojas Stability of Somatoform Symptoms Implications for Classification Psychosom Med, November 1, 2007; 69(9): 864 - 869. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. J. Brown, E. Cardena, E. Nijenhuis, V. Sar, and O. van der Hart Should Conversion Disorder Be Reclassified as a Dissociative Disorder in DSM V? Psychosomatics, October 1, 2007; 48(5): 369 - 378. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Waitzkin, J. Yager, T. Parker, and B. Duran Mentoring Partnerships for Minority Faculty and Graduate Students in Mental Health Services Research Acad Psychiatry, June 1, 2006; 30(3): 205 - 217. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Van Diest, S. De Peuter, K. Piedfort, J. Bresseleers, S. Devriese, K. P. Van de Woestijne, and O. Van den Bergh Acquired lightheadedness in response to odors after hyperventilation. Psychosom Med, March 1, 2006; 68(2): 340 - 347. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
