Journal of Affective Disorders 78 (2004) 37 – 48 www.elsevier.com/locate/jad

Research report

Partial posttraumatic stress disorder revisited Jacques Mylle a,*, Michael Maes b,c a

Department of Psychology, Royal Military Academy (RMA), Renaissancelaan, 30, B1000 Brussels, Belgium b Clinical Research Center for Mental Health (CRC-MH), Antwerp, Belgium c Department of Psychiatry, University of Maastricht, Maastricht, The Netherlands Received in revised form 14 June 2002; accepted 15 June 2002

Abstract Background: It is thought that the decision rule for a positive diagnosis of Posttraumatic Stress Disorder (PTSD) may be too restrictive, leaving too many victims of a trauma out in the cold for care, compensation, etc. Several authors have proposed the concept of Subthreshold or Partial PTSD (PPTSD). This concept considers that a subject may present a number of symptoms below threshold for criteria C or D (subthreshold syndromes) and may even present without any symptom for one or more of the criteria B, C and D (partial syndromes). Method: Data have been collected by means of the Composite International Diagnostic Interview (CIDI) PTSD-module, in a group exposed to two different traumatic events (130 fire victims and 55 car accident victims). The syndrome patterns has been assessed by means of hierarchical class analyses. Each of the criteria B, C and D has been analyzed separately, showing the symptom patterns as hierarchically order clusters. Results: Depending on the threshold used for criterion C (i.e. 3 or 2 symptoms), 18.4 and 22.7% of the subjects respectively satisfy the criteria for PTSD. 8.7% of the subjects show subthreshold syndromes. 60.7% of the subjects show partial syndromes and 16.7% of the subjects have partial syndromes while fulfilling criterion F, i.e. a clinically significant impairment in functioning. Conclusions: The results show a considerable number of partial and subthreshold syndromes. It is argued that subthreshold syndromes and partial syndromes, which fulfill criterion F, should be regarded as specific nosological categories or as specified PTSD subcategories, i.e. subsyndromal or partial PTSD. D 2002 Elsevier B.V. All rights reserved. Keywords: Posttraumatic stress disorder; Hierarchical classes analysis; Burn; Motor vehicle accidents

1. Introduction Since Posttraumatic Stress Disorder (PTSD) has been introduced as a nosological category in the

* Corresponding author. Tel.: +32-2-737-6600; fax: +32-2-7376512. E-mail address: [email protected] (J. Mylle). 0165-0327/$ - see front matter D 2002 Elsevier B.V. All rights reserved. doi:10.1016/S0165-0327(02)00218-5

Diagnostic and Statistical Manual for Mental Disorders, third edition (DSM-III; APA, 1980), research has been undertaken concerning the epidemiology of PTSD and the validity of the diagnostic criteria as well. Despite a revised version, DSM-IIIR (APA, 1987), and a new edition, DSM-IV (APA, 1994) the concept PTSD is still a matter of debate. One issue deals with the threshold for a criterion in terms of a minimal number of symptoms to be

38

J. Mylle, M. Maes / Journal of Affective Disorders 78 (2004) 37–48

present for a positive diagnosis. According to DSMIII-R and DSM-IV, PTSD is defined as a syndrome, following a traumatic event (criterion A), of, at least, one intrusion symptom (criterion B), three avoidance and/or numbing symptoms (criterion C), and two hyperarousal symptoms (criterion D). These symptoms have to persist for at least 1 month (criterion E). According to DSM-IV, their intensity must be such that they cause a clinical significant impairment in functioning of the victim in at least one important domain of life (criterion F). This decision rule may be too restrictive since a number of victims do not fulfill the DSM criteria, although they may suffer from disturbing symptoms of PTSD. In order to solve this problem, some authors (Blank, 1992; Carlier and Gersons, 1995; Kulka et al., 1988; Parson, 1990, Weiss et al., 1992) have proposed the concepts Partial PTSD (PPTSD), subthreshold PTSD or subsyndromal PTSD. Although high prevalences of these syndromes have been reported, this concept has not been accepted by the DSM IV. Adjustment disorder is the DSMIV nearest equivalent of these conditions. One of the ways of defining an adjustment disorder is that it is a condition with a sub-threshold traumatic event (criterion A) or a sub-threshold response (B to E). The prevalence of PPTSD in the population, however, may be different between studies depending on the different operationalizations. Reviewing the literature, two main profiles may be found. The first one, assumes that at least one symptom is present for each of the criteria B, C and D. For example, Kulka et al. (1988) consider that at least two symptoms belonging to criterion C should be present. The second PPTSD profile considers that a combination of two of the three criteria B, C and D should be present. It must be stressed that it is not always clear whether the criteria E and F still apply to the ‘‘diagnosis’’ PPTSD. For example, Blanchard et al. (1994) consider the forms BC and BD. Carlier and Gersons (1995) use seven classes of PPTSD: B, C, D, BC, BD, CD and one unspecified for a clinical image that is composed of a number of symptoms which do not satisfy any criterion (e.g. two symptoms C and one symptom D). Using two widespread databases, PsycLIT and Medline, we found 26 references dealing with

PPTSD, 14 of them referring to a trauma following a car accident or severe burn injuries. Table 1 gives an overview of these 14 studies. Depending on the definition of PPTSD, the current incidence ranges from 2.8% following a car accident (Malt, 1988) to 86.5% among severely burned people (McLeer et al., 1992). In the latter, however, criterion B is the only one being satisfied. The lifetime prevalence in American population studies is estimated between 6.6% (Davidson et al., 1991) and 15% (Helzer et al., 1987). The aims of the present study are to determine: (1) the various clinical patterns of PPTSD that may occur in a group of victims who have experienced a blaze or a severe car pile up with fire; (2) the incidence of these patterns; and (3) the changes in the incidence rate of PPTSD when the threshold for criterion C is reduced from 3 to 2 symptoms.

2. Method 2.1. Collection of data The subjects are victims of two traumatizing events. On December 31, 1994, around midnight a blaze destroyed the ballroom of the Switel Hotel (Antwerp, Belgium), in which about 450 guests were celebrating New Years Eve. More than 120 subjects were injured or burned; ten people died on scene or later. On February 27, 1996, around 09.00 A.M., on the E17 highway (at Deinze, Belgium) a sudden heavy fog bank surprised several drivers, most driving too close to each other, resulting in a pile-up with more than 150 vehicles involved. Many cars were crushed with explosions and fire as a consequence. Ten persons died at scene and more than 50 were injured. Asked for their participation in a research project on PTSD, 130 fire victims and 55 pile up victims were willing to complete structured interviews. All subjects are White, most of them are Flemish and some are Dutch. Three trained research assistants carried out the interviews 7 to 9 months after the events. The Composite International Diagnostic Interview, PTSD module (CIDI-PTSD) was used in a validated Dutch translation (Smeets and Dingemans, 1993) assessing PTSD symptoms according to the DSM-III-R crite-

J. Mylle, M. Maes / Journal of Affective Disorders 78 (2004) 37–48

39

Table 1 Overview of studies reporting PPTSD following a MVA or a fire disaster No

Authors

Subjects

1

Malt (1988)

Motor vehicle accident

2 3 4 5

Hickling and Blanchard (1992) Hickling et al. (1992) Kuch et al. (1994) Blanchard et al. (1994)

Motor vehicle accident Motor vehicle accident Motor vehicle accident. Accidents

6

Blanchard et al. (1995)

Form

7 8 9 10

Chibnall and Duckro (1994) Taylor and Koch (1995) Carlier and Gersons (1995) Green et al. (1993)

11

McLeer et al. (1988)

Accidents (T + 1 to 4 months) (T + 6 months) Motor vehicle accident + headaches Accidents Airplane crash on buildings Inpatient victims (T + 3 months) (T + 18 months) Heavy burn injuries (children)

12 13

Stoddard et al. (1989) Patterson et al. (1990)

Heavy burn injuries (children + adolescents) Heavy burn injuries

14

McLeer et al. (1992)

Heavy burn injuries

? B C ? ? ? Part BC/BD id ? ? Part ? Part B C D ? Part B Part B C D

Prevalence (%) PTSD

PPTSD

0.93

2.80 21.0 44.0 5 – 15 id id

38 – 75 id id 46.0 41.0 21.0 290 42.0 26.0 8.0 30.0 48.4

6.7 29.6

20.0 26.0 22.0 20.0 27.0 44.0 27 5.0 ? 80.6 48.3 64.5 20 63.0

43.9

? 86.5 52.4 72.0

Sub: subsyndromal form; Part: partial form; ?: unspecified, B, C, D: PTSD considered.

ria. The items are scored in a dichotomous way; i.e. the person has or he has not the symptom. 2.2. Statistics We assessed the structure in our data by means of a clustering technique, called hierarchical classes analysis (Hiclas) (De Boeck and Rosenberg, 1988; De Boeck et al., 1992). The starting point for hierarchical classes analysis is a binary object-byattribute matrix. In our study, the objects are the subjects and the attributes are the PTSD criteria, the symptoms per criterion, respectively. The aims of Hiclas are: (1) to uncover sets of objects that share the same class of attributes (and vice versa); and (2) to represent them in a hierarchical way; i.e. in terms of superordinate – subordinate relationships. This means that a lower order set is a subset of a higher

order set. For example, if X, Y and Z are sets of some attributes, then the class XY is a superordinate set of the subsets X and Y, but in turn XY is a subset of the class XYZ. Hiclas offers a display of the results whereby classes are represented as boxes and the hierarchical relationship is reflected by drawing the classes that contain the same number of elements in the same layer. From the bottom to the top layers are drawn according to the increasing number of elements in a set. The union of the lower order sets in a higher order set is expressed by means of lines connecting the classes involved. The graph must be read from the bottom classes in the direction of the top class. Finally, the hierarchies of classes of objects and of classes of attributes are connected together. A special class is the null class that contains objects without any attributes; this class is represented at the very bottom of the visual

40

J. Mylle, M. Maes / Journal of Affective Disorders 78 (2004) 37–48

representation. Classes that are empty—i.e. not containing any items are not represented in the graph. For example, in Fig. 1, the null class contains the subjects who do not satisfy any criterion; the lowest meaningful classes are the ones containing a single criterion (therefore they are called ‘‘bottom classes’’); at the second level appear the classes that are unions of two single attributes, at the third level those with three attributes, and so on. We used as many bottom classes as there are single attributes. By doing so we obtained a perfect solution; this means that all relationships between the objects and the attributes in all classes are correctly allocated. To facilitate the interpretation of the figures with regard to the criteria, we use grey-shaded boxes if criterion F is satisfied. When the threshold for criterion C and/or D is not reached, the code for that particular criterion is placed between brackets. For example, B(C)D means that the threshold for criteria B and D is reached but not for criterion C; although there is at least one symptom C present.

3. Results 3.1. In terms of PTSD criteria First, we computed a solution with the five criteria as defined in DSM-III-R or DSM-IV. We found 20 classes out of the 32 ( = 25) possible patterns represented, but only 18 of them are populated (Fig. 1). Starting at the bottom of Fig. 1, we see that 11.9% belong to the null class, but in fact, this group is composed of 2.7% of the respondents without any symptom, and 9.2% with one or more symptoms below threshold for the criteria C and/or D. At the level of the bottom classes, we see that avoidance/ numbing behavior (C) alone does not exist: the class 10000 is empty. Intrusion (B) in class 00001 and hyperarousal (D) in class 01000 appear separately, with prevalences of 5.4 and 4.9%, respectively. These groups can be split up in two subgroups each, with 1.6 and 2.7% ‘‘pure’’ cases on the one hand, and with 3.8 and 2.2% cases which combine a satisfied criterion with one or two subthreshold criteria on the other; e.g. (C)D or B(CD). It must be noticed that the one person (0.5%) in class 00100 contends to be disturbed in some area of his functioning although he shows only a

subthreshold pattern on C and D (actually one on each). At the second level, classes appear with pairwise combined criteria that are satisfied, completed or not with subthreshold criteria. The most striking group, in terms of prevalence, is the class 01001 with the pattern BD. This group of 16.7% subjects is equally split in 8.6% ‘‘pure’’ BD cases and 8.1% of the B(C)D type. Notwithstanding their symptom pattern, they feel unimpaired in their functioning (F not satisfied). On the contrary, the 3.2% of type (C)DF (class 01100) and the 1.6% with pattern B(C)F c.q. B(D)F (class 00101) may suggest that one criterion may suffice to be disturbed in functioning, even if these symptoms do not persist for longer than 1 month. At the third level, two classes contain a rather large subgroup: class 01011 has 9.7% subjects with a profile B(C)DE and class 01101 has 3.1% BDF and 5.9% B(C)DF, respectively, or 9.1% in total. Although the symptoms persist for more than 1 month in the first group, they feel not impaired, while the latter group reports a clinically significant disturbance in functioning in the short period that their symptoms are present (less than 1 month). At the next higher level, all possible combinations are present except CDEF. The class 01111 is the ‘‘classic’’ type of subthreshold PTSD with the pattern B(C)DEF. This class shows also the highest prevalence at this level, namely 7.6%. The subthreshold pattern BC(D)EF is present but it only with 1.1% of the respondents. The class 11011 with 3.8% BCDE shows that, although all criteria are satisfied, except F, these persons feel not impaired in their functioning and, consequently, are not diagnosed as PTSD cases. Conversely, class 11101 contains 2.2% of the subjects that satisfy all criteria, except criterion E. Finally, the hierarchical highest class contains the 18.4% subjects with a ‘‘full blown PTSD’’. In summary, we have 2.7% cases without symptoms, and 9.2% with only a few below any threshold, 60.7% partial syndromes (composed of 16.7% satisfying F), 8.7% subthreshold syndromes and 18.4% full PTSD cases. 3.2. Impact of lowering the threshold to Cz2 We analyzed the same data with a threshold for criterion C equal to two symptoms as suggested by some authors. Now, 22 patterns out of the 32 possible

J. Mylle, M. Maes / Journal of Affective Disorders 78 (2004) 37–48

Fig. 1. Hiclas representation of PPTSD in terms of criteria with threshold C z 3.

41

42

J. Mylle, M. Maes / Journal of Affective Disorders 78 (2004) 37–48

appear instead of 20 (Fig. 2). The added classes are the bottom class 10000 with pattern C(D), and the third level class 10101 with CDF. In Fig. 2, they are marked with a bold frame. Furthermore, a number of subjects shift from one class into another at a higher level because now they satisfy one criterion more. In Fig. 2, we visualized this phenomenon by marking the box they left by a downward arrow and the box they moved to with an upward arrow, and by putting the line connecting these classes in bold. The number of subjects in the null class is reduced by one, namely the subject that changes from (CD) to C(D) and therefore moves to the new bottom class 10000. Further, two patterns (1.1%) change from (C)D into CD and leave thus the bottom class 00001 for the second level class 10001, which prevalence becomes now 1.6%. The same holds for 1.6% that change B(C) into BC. At the second level, 1.1% (C)DF move into the new third level class CDF; 2.2% B(C)D change into BCD and shift thus to the third level class 11001, that contains now 3.8% of the subjects (instead of 1.6%). At the third level, 3.2% of the respondents show a pattern that turns from B(C)DF into BCDF, bringing the prevalence of this fourth level class to 5.4%. The most important change however concerns the 5.9% B(C)DE that move into the BCDE class, which prevalence is now 9.7%. Finally, the relaxation of the criterion results in the change of 4.4% B(C)DEF into PTSD diagnosis; there are now 22.7% full blown PTSD cases. Taken together, there are 2.7% cases without any symptom, and 8.7% with a number of symptoms below any threshold. There are 61.3% partial syndromes (44.6% not satisfying F and 16.7% who do). The prevalence of subsyndromal cases is 4.3%, while 22.7% subjects receive a positive PTSD diagnosis. It should be noticed that lowering the threshold for C does not have an impact on the prevalence of the other criteria; thus the number of cases that feel impaired in their functioning due to their syndrome remains the same as under the former condition; i.e. 43.8% satisfy criterion F and 53.6% do not. 3.3. Clusters of symptoms within criteria The four symptoms belonging to criterion B described in DSM-III-R appear in all possible combinations; in other words the 24 = 16 patterns exist in

the target group. The structure and the prevalences of each class is shown in Fig. 3. It should be noticed that the prevalence is the highest in those classes in which ‘‘intrusive recollections’’ (I) are present. At the level of the bottom classes, the prevalence for I is 10.8% whereas for the other three symptoms it is only 1.6% for dissociation (D), 2.2% for nightmares (N) and 3.8% for emotional reactions (E). At the second level, the three pairs in which I appears—i.e. IN, ID, and IE—amount to 28%, while the other three together total only 4.3%. Among the former, ‘‘intrusive recollections – emotional reactions’’ (IE) accounts for 12.4% and ‘‘intrusive recollections – nightmares’’ (IN) for 9.7%. Even the combination ‘‘intrusive recollections – dissociation’’ (ID) (5.9%) has a higher prevalence than the other three. This trend holds on at the level of the triplets: the prevalence of ‘‘intrusive recollections—nightmares– emotional reactions’’ (INE) is 7% while the other are 4.3, 3.8 and 2.2.% only. Finally, 9.7% show all four symptoms. The seven symptoms of criterion C can be combined in 128 patterns out of which 61 are observed. The first three symptoms in Fig. 4 refer to ‘‘avoidance behavior’’ while the four next are ‘‘psychic numbing’’ symptoms. The null class groups 28.1% of the subjects. Given that the number of classes is relatively large compared to the number of subjects (61:185) the prevalence of most of the classes is low. The prevalence for the successive levels is 24.3, 19, 10.3, 7, 8.1 1.1 and finally 1.6%. Among the bottom classes, those referring to avoidance behavior show higher prevalences than those referring to psychic numbing, i.e. 19.7 versus 4.8%, respectively. With respect to avoidance behavior, 9.7% of the subjects have difficulties to remember important aspects of the trauma (R), 6.5% ‘‘avoid activities, situations and persons that remind of the trauma’’ (A) and 3.2% ‘‘avoid to think of the trauma’’ (T). With respect to numbing behavior, 3.2% of the subjects show ‘‘diminished interest’’ (I) and 1.6% ‘‘feelings of detachment’’ (D). ‘‘Restriction in expression of affect’’ (E) and ‘‘foreshortened future’’ (F) do not exist as single symptoms. At the second level, the three pairs involving two avoidance symptoms outnumber by and large those involving psychic numbing with prevalence of 6.5% (TA), 3.8% (AR) and 2.2% (TR), respectively, whereas the other classes are limited to

J. Mylle, M. Maes / Journal of Affective Disorders 78 (2004) 37–48

Fig. 2. Hiclas representation of PPTSD in terms of criteria with threshold C z 2.

43

44

J. Mylle, M. Maes / Journal of Affective Disorders 78 (2004) 37–48

Fig. 3. Hiclas representation of the symptoms patterns—criterion B.

0.5 or 1.1%. All classes of all higher levels are sparsely populated with 0.5 or 1.1%, except for the class AIDEF, combining all numbing symptoms with avoidance of activities, which prevalence is 2.2%. Remarkably, at the subtop level only two of the seven possible classes are effectively present; be it with very small a prevalence (0.5%). Finally, 1.7% report all seven symptoms. Criterion D is defined by six symptoms in DSMIII-R. Forty-five classes out of the 64 possible appear in the structure (Fig. 5). Only 8.1% of the subjects do not show any symptoms. Among the bottom classes, the class ‘‘hypervigilance’’ (H) contains most subjects (8.6%) while ‘‘increased irritability’’ (I) and ‘‘physiological reactivity’’ (P) do not appear as single symptoms. At the second level, the pair ‘‘sleep disorder –hyper vigilance’’ (SH) has the highest prevalence (4.9%) followed by ‘‘hypervigilance –exaggerated startles’’ (HE) with 3.9%. The same holds at the level of the triplets; namely ‘‘sleep disorder –hypervigilance –physiological reactivity’’ (SHP) with 3.8% and ‘‘sleep disorder – hypervigilance – exaggerated startles’’ (SHE) with 3.2%. At the fourth level, the class containing the aforementioned symptoms (SHEP) is most populated. At the subtop level, the classes with ‘‘sleep disorder’’ (S), ‘‘concentration difficulties’’ (C) and ‘‘hypervigilance’’ (H) are best represented with 3.2, 3.8 and 3.8%, respectively. Finally, 9.2% show all six symptoms.

4. Discussion In the present study, we found that a considerable number of victims showed subthreshold (8.7%) or partial (60.7%) syndromes. From a clinical point of view, it seems reasonable to accept only those partial syndromes for which the criterion F is satisfied because only these victims may need professional treatment. 16.7% of all subjects suffered from partial syndromes and at the same time fulfilled the F criterion, although none of these forms satisfies the time criterion E. Thus, the latter group reports a partial syndromes and a clinically significant disturbance in psycho-social functioning in the shorter period that their symptoms are present, i.e. less than 1 month. Using the criteria for subthreshold syndromes and those partial syndromes which fulfill the F criterion rises the number of cases in the target group (PTSD and subthreshold/partial syndromes) with 8.7 and 16.7%, respectively. Given the wide variety of decision rules that have been used to evaluate the prevalence of PPTSD, it is difficult to compare the results reported in Table 1 with ours. Thus, the results vary from 2.80% (Malt, 1988) to 44% (Carlier and Gersons, 1995) for MVA, and between 48% (only criterion C present; McLeer et al., 1988) and 86.5% (only criterion B present; McLeer et al., 1992) in the case of fire burn victims. Although several authors report PPTSD forms like

J. Mylle, M. Maes / Journal of Affective Disorders 78 (2004) 37–48

Fig. 4. Hiclas representation of the symptoms patterns—criterion C.

45

46

J. Mylle, M. Maes / Journal of Affective Disorders 78 (2004) 37–48

Fig. 5. Hiclas representation of the symptoms patterns—criterion D.

BDF c.q. BCF, and even BF, CF, DF the question remains open if these patterns are typical for posttraumatic stress or arise also in other conditions. There is some evidence that subthreshold or partial PTSD syndromes may differ from PTSD. First, PPTSD has typical symptom patterns, e.g. BD and BC, discriminating it from PTSD. Second, Van Driel and Op den Velde (1995) have shown that PTSD and PPTSD cases may use different defense mechanisms. For example, both use projection, somatization and conversion, but the PPTSD victims use more repression, denial and splitting. The phenomenological description of subsyndromal and partial PTSD, which has been delineated in the present study, should be validated against external validating criteria, such as neurobiological markers (Maes et al., 1998, 1999). One question is whether these subsyndromes of PTSD have the same biological patho-

physiology of PTSD and can be distinguished from adjustment disorder or simple stress reactions. In addition, another key distinction should be made between: (a) those reactions which follow a severe trauma (meeting the criterion in full in PTSD) but which do not have all the symptoms for the diagnosis; and b) those reactions which may have similar symptoms but which follow a lesser event. This distinction is emphasized in the DSM-IV in that in adjustment disorder personal vulnerability is often seen as more important. Many authors cited in Table 1 underline that the problem of PTSD versus PPTSD has not only a taxonomic utility—PPTSD as a developing PTSD, or as a PTSD in remission, or as a syndrome suis generis—but has also important consequences for clinicians, public health policy makers, public health workers, and, last but not least, for the trauma

J. Mylle, M. Maes / Journal of Affective Disorders 78 (2004) 37–48

victims themselves. Indeed, a nosologic taxonomy, such as the DSM also defines the medico-legal position of the subjects. By inference, it is very important to decide whether victims with subthreshold PTSD or partial PTSD should be considered to have a same medico-social position as PTSD patients. In any case, since our study and some of the studies reviewed in Table 1 show that the prevalence of PPTSD may be very high, we suggest that subthreshold syndromes and partial syndromes, which fulfill criterion F, should be regarded as specific nosological categories or as specified PTSD subcategories. Therefore, we propose to label: (1) ‘‘subsyndromal PTSD’’ those syndromes in which the threshold for criterion C and/or D is not reached (subthreshold syndromes) but with at least one symptom of each criterion present; and (2) ‘‘Partial PTSD’’ when the symptoms for a certain criterion are absent, but when criterion F is present.

Acknowledgements The research reported was supported in part by the Fund for Scientific Research, Vlaanderen (FWO); the Clinical Research Center for Mental Health (CRCMH), Antwerp; The Ministry of Welfare, Belgium; Pfizer-Belgium and Belgacom, Belgium. The assistance of L. Delmeire, M.D., E. De Boel; G. Belis; C. Schotte, and A. Van Gastel, M.D., CRC-MH, Antwerp; Prof. B. Van Houdenhove, M.D., KUL, Leuven, Belgium; and C.S. North, M.D., Associate Professor, Washington University, St. Louis, USA is greatly appreciated. The authors would like to thank Slachtofferhulp Vlaanderen; the ‘‘Stad Antwerpen’’ (L. Detiege); Politie van Antwerpen and Slachtofferzorg (B. Engelen, R. Van Looveren); Slachtofferhulp, Antwerpen; and the Red Cross, Vlaanderen and the ‘‘Dienst Dringende Spoed Interventie’’ (D. De Beukelaer, H. Van Gastel).

References American Psychiatric Association, 1980. Diagnostic and Statistical Manual of Mental Disorders, 3th Edition. APA, Washington, DC. American Psychiatric Association, 1987. Diagnostic and Statistical

47

Manual of Mental Disorders, (3th Edition, rev). APA, Washington, DC. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. APA, Washington, DC. Blanchard, E.B., Hickling, E.J., Taylor, A.E., Loos, W.R., Gerardi, R.J., 1994. Psychological morbidity associated with motor vehicle accidents. Behav. Res. Ther. 32, 283 – 290. Blanchard, E.B., Hickling, E.J., Taylor, A.E., Loos, W.R., 1995. Psychiatric morbidity associated with motor vehicle accidents. J. Nerv. Ment. Dis. 182 (8), 495 – 504. Blank Jr., A.S., 1992. The longitudinal course of posttraumatic stress disorder. In: Davidson, J.R.T., Foa, E.B. (Eds.), Posttraumatic Stress Disorder: DSM-IV and Beyond. American Psychiatric Press, Washington, DC. Carlier, I.V., Gersons, B.P., 1995. Partial Posttraumatic Stress Disorder (PTSD): the issue of psychological scars and the occurrence of PTSD symptoms. J. Nerv. Ment. Dis. 183 (2), 107 – 109. Chibnall .Y., J.Y., Duckro, P.N., 1994. Post-traumatic stress disorder in chronic posttraumatic headache patients. Headache 34, 357 – 361. Davidson, J.R.T., Hughes, D., Blazer, D. et al., 1991. Post traumatic stress disorder in the community: an epidemiological study. Psychol. Med. 21, 1 – 9. De Boeck, P., Rosenberg, S., 1988. Hierarchical classes: Model and data analysis. Psychometrika 53, 361 – 381. De Boeck, P., Van Damme, G., Van Mechelen, I., 1992. HICLAS computer program: Version 2.0. Leuven, Katholieke Universiteit. Green, M.M., McFarlane, A.C., Hunter, C.E., Criggs, W.M., 1993. Undiagnosed posttraumatic stress disorder following motor accidents. Med. J. Austr. 159, 529 – 534. Helzer, J.E., Robins, L.N., McEvoy, L., 1987. Posttraumatic stress disorder in the general population. N. Engl. J. Med. 317, 1630 – 1634. Hickling, E.J., Blanchard, E.B., 1992. Posttraumatic stress disorder and motor vehicle accidents. J. Anxiety Disord. 6, 285 – 291. Hickling, E.J., Blanchard, E.B., Silverman, D.J., Schwartz, S.P., 1992. Motor vehicle accidents, headaches and posttraumatic stress disorder. Assessment findings in a consecutive series. Headache 32, 147 – 151. Kuch, K., Cox, B.J., Evans, R.J., Shulman, I., 1994. Phobias, panic, and pain in 55 survivors of road vehicle accidents. J. Anxiety Disord. 8, 181 – 187. Kulka, R.A., Schlenger, W.E., Fairbank, J.A., Hough, R.L., Jordan, B.K., Marmar, C.R., Weiss, D.S., 1988. National Vietnam veterans readjustment study advance data report: Preliminary findings from the national survey of the Vietnam generation. Veterans Administration, Washington, DC. Malt, U., 1988. The long-term psychiatric consequences of accidental injury: A longitudinal study of 107 adults. Br. J. Psychiatry 153, 810 – 818. McLeer, S.V., Deblinger, E., Atkins, M.S., Foa, E.B. et al., 1988. Posttraumatic stress disorder in sexually abused children. Annual Meeting of the Academy of Child and Adolescent Psychiatry, 1987, Washington, DC. J. Am. Acad. Child Adol. Psychiatry 27 (5), 650 – 654.

48

J. Mylle, M. Maes / Journal of Affective Disorders 78 (2004) 37–48

McLeer, S.V., Deblinger, E., Henry, D., Orvaschel, H., 1992. Sexually abused children at high risk for post-traumatic stress disorder. J. Am. Acad. Child Adol. Psychiatry 31 (5), 875 – 879. Maes, M., Lin, A., Bonaccorso, S., van hunsel, F., Van Gastel, A., Delmeire, L., Biondi, M., Bosmans, E., Kenis, G., Scharpe, S., 1998. Increased baseline hypothalamic – pituitary – adrenal axis activity in patients with posttraumatic stress disorder and major depression, but not in patients with fibromyalgia. Acta Psychiatr. Scand. 98, 328 – 335. Maes, M., Lin, A., Verkerk, R., Delmeire, L., Van Gastel, A., Gommeren, W., Leysen, J., Scharpe, S., 1999. Serotonergic and noradrenergic markers of posttraumatic stress disorder: lower affinity of platelet paroxetine and alpha-2 binding sites and increased plasma tyrosine availability to the brain. Neuropsychopharmacology 20, 188 – 197. Parson, E.R., 1990. Post traumatic demoralization syndrome. J. Contemp. Psychother. 20 (1), 17 – 33. Patterson, D.R., Carrigan, L., Questad, K.A., Robinson, R., 1990.

Post-traumatic stress disorder in hospitalized patients with burn injuries. J. Burn Care Rehab. 11 (3), 181 – 184. Smeets, M.R., Dingemans, P.M., 1993. Composite International Diagnostic Interview, Version 1.1, Interviewers Version, Dutch translation, World Health Organization. Stoddard, F.J., Norman, D.K., Murphy, J.M., 1989. A diagnostic outcome study of children and adolescents with severe burns. J. Trauma 29 (4), 471 – 477. Taylor, S., Koch, W., 1995. Anxiety disorders due to motor vehicle accidents: Nature and treatment. Clin. Psychol. Rev. 15, 721 – 738. Van Driel, R.C., Op den Velde, W., 1995. Myocardial infarction and posttraumatic stress disorder. J. Trauma Stress 8, 151 – 159. Weiss, D.S., Marmar, C.R., Schlenger, W.E., Fairbank, J.A. et al., 1992. The prevalence of lifetime and partial post-traumatic stress disorder in Vietnam theater veterans. J. Trauma Stress 5, 365 – 376.