This systematic review and collaborative recalculation was set up to recalculate schizophrenia incidence rates from previously published studies by age and sex.

PubMed, EMBASE and PsycINFO databases were searched (January 1950 to December 2009) for schizophrenia incidence studies. Numerator and population data were extracted by age, sex and, if possible, study period. Original data were requested from the authors to calculate age- and sex-specific incidence rates. Incidence rate ratios (IRRs) with their 95% confidence intervals (CIs) were computed by age and sex from negative binomial regression models.

Forty-three independent samples met inclusion criteria, yielding 133 693 incident cases of schizophrenia for analysis. Men had a 1.15-fold (95% CI 1.00–1.31) greater risk of schizophrenia than women. In men, incidence peaked at age 20–29 years (median rate 4.15/10 000 person-years, IRR 2.61, 95% CI 1.74–3.92). In women, incidence peaked at age 20–29 (median rate 1.71/10 000 person-years, IRR 2.34, 95% CI 1.66–3.28) and 30–39 years (median rate 1.24/10 000 person-years, IRR 2.25, 95% CI 1.55–3.28). This peak was followed by an age–incidence decline up to age 60 years that was stronger in men than in women (χ2 = 57.90, p < 0.001). The relative risk of schizophrenia was greater in men up to age 39 years and this reversed to a greater relative risk in women over the age groups 50–70 years. No evidence for a second incidence peak in middle-aged women was found.

Robust sex differences exist in the distribution of schizophrenia risk across the age span, suggesting differential susceptibility to schizophrenia for men and women at different stages of life.

The past two decades have seen exponential clinical and research interest in help-seeking individuals presenting with potentially prodromal symptoms for psychosis. However, the epidemiological validity of this paradigm has been neglected, limiting future advancements in the field.

We undertook a critical review of core epidemiological issues underlying the clinical high-risk (HR) state for psychosis and which model of prodromal intervention is best suited for mental health.

The HR state for psychosis model needs refining, to bring together population-based findings of high levels of psychotic experiences (PEs) and clinical expression of risk. Traditionally, outcome has been attributed to ‘HR criteria’ alone rather than taking into account sampling strategies. Furthermore, the exclusive focus on variably defined ‘transition’ obscures true variation in the slow and non-linear progression across stages of psychopathology. Finally, the outcome from HR states is variable, indicating that the underlying paradigm of ‘schizophrenia light progressing to schizophrenia’ is inadequate.

In the general population, mixed and non-specific expression of psychosis, depression, anxiety and subthreshold mania is common and mostly transitory. When combined with distress, it may be considered as the first, diagnostically neutral stage of potentially more severe psychopathology, which only later may acquire a degree of diagnostic specificity and possible relative resistance to treatment. Therefore, rather than creating silos of per-disorder ultra-HR syndromes, an early intervention focus on the broad syndrome of early mental distress, requiring phase-specific interventions, may be more profitable.

Intact sarcasm perception is a crucial component of social cognition and mentalizing (the ability to understand the mental state of oneself and others). In sarcasm, tone of voice is used to negate the literal meaning of an utterance. In particular, changes in pitch are used to distinguish between sincere and sarcastic utterances. Schizophrenia patients show well-replicated deficits in auditory function and functional connectivity (FC) within and between auditory cortical regions. In this study we investigated the contributions of auditory deficits to sarcasm perception in schizophrenia.

Auditory measures including pitch processing, auditory emotion recognition (AER) and sarcasm detection were obtained from 76 patients with schizophrenia/schizo-affective disorder and 72 controls. Resting-state FC (rsFC) was obtained from a subsample and was analyzed using seeds placed in both auditory cortex and meta-analysis-defined core-mentalizing regions relative to auditory performance.

Patients showed large effect-size deficits across auditory measures. Sarcasm deficits correlated significantly with general functioning and impaired pitch processing both across groups and within the patient group alone. Patients also showed reduced sensitivity to alterations in mean pitch and variability. For patients, sarcasm discrimination correlated exclusively with the level of rsFC within primary auditory regions whereas for controls, correlations were observed exclusively within core-mentalizing regions (the right posterior superior temporal gyrus, anterior superior temporal sulcus and insula, and left posterior medial temporal gyrus).

These findings confirm the contribution of auditory deficits to theory of mind (ToM) impairments in schizophrenia, and demonstrate that FC within auditory, but not core-mentalizing, regions is rate limiting with respect to sarcasm detection in schizophrenia.

Trajectory patterns of positive, disorganized and negative dimension symptoms during antipsychotic treatment in drug-naive patients with first-episode psychosis have yet to be examined by using naturalistic data.

This pragmatic clinical trial randomized 161 drug-naive patients with a first episode of psychosis to olanzapine, risperidone or haloperidol. Patients were assessed with the Scale for the Assessment of Negative Symptoms (SANS) and Positive Symptoms (SAPS) at baseline and at the end of weeks 1, 2, 3, 4 and 6 of antipsychotic treatment. Censored normal models of response trajectories were developed with three dimensions of the SAPS-SANS scores (positive, disorganized and negative) in order to identify the different response trajectories. Diagnosis, cannabis use, duration of untreated psychosis (DUP), smoking and antipsychotic class were examined as possible predictive variables.

Patients were classified in five groups according to the positive dimension, three groups according to the disorganized dimension and five groups according to the negative dimension. Longer DUPs and cannabis use were associated with higher scores and poorer responses in the positive dimension. Cannabis use was associated with higher scores and poorer responses in the disorganized dimension. Only schizophrenia diagnosis was associated with higher scores and poorer responses in the negative dimension.

Our results illustrate the heterogeneity of short-term response to antipsychotics in patients with a first episode of psychosis and highlight markedly different patterns of response in the positive, disorganized and negative dimensions. DUP, cannabis use and diagnosis appeared to have a prognostic value in predicting treatment response with different implications for each dimension.

Few studies have examined the latent construct of psychotic symptoms or distinguished between the latent construct and its manifest indicators. The current study aimed to investigate the latent structure of psychotic symptoms using factor mixture modeling (FMM) and to use the best-fitting model to examine its sociodemographic and clinical correlates.

The Singapore Mental Health Study (SMHS) was based on an adult representative sample of the Singapore population. Psychotic symptoms were assessed by using the Psychosis Screen section of the Composite International Diagnostic Interview version 3.0 (CIDI 3.0). FMM analyses were applied to determine the latent construct of psychotic symptoms. Sociodemographic and clinical correlates of the latent structure of psychosis symptoms were examined using multiple linear and logistic regression analyses.

The overall weighted lifetime prevalence of any psychotic experience was 3.8% in the SMHS after excluding subthreshold experiences. The FMM analysis clearly supported the dimensional model of the latent structure of psychotic symptoms. On deriving the total score for ‘psychosis symptoms’ in accordance with the one latent trait model, and correlating it with sociodemographic factors, we found that female gender, vocational education, current and past smokers were positively associated with the ‘psychosis’ total score.

There is a need for an increased understanding of, and research into, this intermediate state of ‘psychosis symptoms’ that do not meet diagnostic criteria for psychosis. It is also important to learn more about the group of individuals in the community who may have preserved functioning to elucidate the protective factors that prevent transition to psychosis.

Illicit drug use is common in individuals with schizophrenia, and it has been suspected that many individuals under-report their use of substances, leading to significant barriers to treatment. This study sought to examine the degree to which individuals with schizophrenia disclose their use of drugs on self-rated assessments, compared to laboratory assays, and to determine the contributors of under-reported drug use in this population.

A total of 1042 individuals with schizophrenia who participated in screening/baseline procedures for the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) completed self-rated assessments of substance use and laboratory drug testing. Laboratory tests assayed cannabis, cocaine and methamphetamine use; the procedures included radioimmunoassay (RIA) and urine drug screens.

A significant proportion of participants tested positive for drug use on laboratory measures (n = 397; 38%), and more than half (n = 229; 58%) did not report using these drugs. Logistic regression models confirmed that patients who were most likely to conceal their use tended to be older, and presented with greater neurocognitive deficits. Patients who accurately reported drug use tended to have greater involvement with the criminal justice system. Illness severity and psychopathology were not associated with whether patients disclosed drug use.

Rates of under-reported drug use are considerable among individuals with schizophrenia when compared to laboratory assays, and the exclusive reliance on self-rated assessments should be used with caution. Patients who under-report their drug use are more likely to manifest neurocognitive deficits, which could be improved by interventions attempting to optimize treatment.

Preconception, prenatal and postnatal maternal stress is associated with increased offspring psychopathology, but findings are inconsistent and need replication. We estimated associations between maternal bereavement stress and offspring autism spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD), bipolar disorder, schizophrenia, suicide attempt and completed suicide.

Using Swedish registers, we conducted the largest population-based study to date examining associations between stress exposure in 738 144 offspring born 1992–2000 for childhood outcomes and 2 155 221 offspring born 1973–1997 for adult outcomes with follow-up to 2009. Maternal stress was defined as death of a first-degree relative during (a) the 6 months before conception, (b) pregnancy or (c) the first two postnatal years. Cox proportional survival analyses were used to obtain hazard ratios (HRs) in unadjusted and adjusted analyses.

Marginal increased risk of bipolar disorder and schizophrenia following preconception bereavement stress was not significant. Third-trimester prenatal stress increased the risk of ASD [adjusted HR (aHR) 1.58, 95% confidence interval (CI) 1.15–2.17] and ADHD (aHR 1.31, 95% CI 1.04–1.66). First postnatal year stress increased the risk of offspring suicide attempt (aHR 1.13, 95% CI 1.02–1.25) and completed suicide (aHR 1.51, 95% CI 1.08–2.11). Bereavement stress during the second postnatal year increased the risk of ASD (aHR 1.30, 95% CI 1.09–1.55).

Further research is needed regarding associations between preconception stress and psychopathological outcomes. Prenatal bereavement stress increases the risk of offspring ASD and ADHD. Postnatal bereavement stress moderately increases the risk of offspring suicide attempt, completed suicide and ASD. Smaller previous studies may have overestimated associations between early stress and psychopathological outcomes.

Emotional dysregulation is becoming increasingly recognized as an important feature of attention deficit hyperactivity disorder (ADHD). In this study, two experiments were conducted investigating the neural response to either verbally instructed fear (IF) or uninstructed (classically conditioned) fear (UF) using the skin conductance response (SCR) and functional magnetic resonance imaging (fMRI).

In the conditioning phase of the UF experiment (17 ADHD and 17 healthy controls), subjects experienced an unconditioned stimulus (UCS, unpleasant electrodermal stimulation) paired with a former neutral conditioned stimulus (CS+), whereas a control stimulus (CS−) was never paired with the UCS. In the subsequent test phase, only the CS+ and the CS− were presented. In the IF experiment (13 ADHD and 17 healthy controls), subjects were only told that an independently experienced UCS might occur together with the CS+ but not the CS− during testing. No UCS was presented.

Groups did not detectably differ in SCR or neural responses to UF. In IF, ADHD patients showed a trend-line decreased SCR and significantly decreased activation of the dorsal anterior cingulate cortex (dACC), a region prominently involved in fear responding, to the CS+. This was accompanied by higher amygdala activation to the CS−.

During IF, ADHD patients showed deficits in regions centrally involved in fear learning and expression in terms of diminished CS+-related dACC and increased CS−-related amygdala signals. This suggests an impaired processing of verbally transmitted aversive information, which is central for conveying fear information in social contexts. This result extends the growing literature on emotional alterations in ADHD.

Children with conduct problems (CP) are a heterogeneous group. Those with high levels of callous–unemotional traits (CP/HCU) appear emotionally under-reactive at behavioural and neural levels whereas those with low levels of CU traits (CP/LCU) appear emotionally over-reactive, compared with typically developing (TD) controls. Investigating the degree to which these patterns of emotional reactivity are malleable may have important translational implications. Instructing participants with CP/HCU to focus on the eyes of fearful faces (i.e. the most salient feature) can ameliorate their fear-recognition deficits, but it is unknown whether this is mediated by amygdala response. It is also unknown whether focusing on fearful eyes is associated with increased amygdala reactivity in CP/LCU.

Functional magnetic resonance imaging (fMRI) was used to measure neural responses to fearful and calm faces in children with CP/HCU, CP/LCU and TD controls (n = 17 per group). On half of trials participants looked for a blue dot anywhere within target faces; on the other half, participants were directed to focus on the eye region.

Reaction time (RT) data showed that CP/LCU were selectively slowed in the fear/eyes condition. For the same condition, CP/LCU also showed increased amygdala and subgenual anterior cingulate cortex (sgACC)/orbitofrontal cortex (OFC) responses compared with TD controls. RT and amygdala response to fear/eyes were correlated in CP/LCU only. No effects of focusing on the eye region were observed in CP/HCU.

These data extend the evidence base suggesting that CU traits index meaningful heterogeneity in conduct problems. Focusing on regulating reactive emotional responses may be a fruitful strategy for children with CP/LCU.

The occurrence of a significant relationship between psychopathic traits and intelligence is still open to debate. Most of the relevant information has been obtained from crystallized IQ tests or on psychopathic male offenders. In this study we hypothesized a negative correlation between psychopathic traits and fluid intelligence on a sample of criminal female in-patients.

We carried out a correlational study on a selected sample of 56 criminal female offenders. Variables that were measured include the Hare Psychopathy Checklist Revised (PCL-R) total score (and, separately, the scores from its four subscales: Interpersonal, Affective, Lifestyle and Antisocial) and fluid IQ measured by Raven's Progressive Matrices (RPM).

Pearson's correlation between RPM IQ and total PCL-R score was negative (r54 = − 0.55, p < 0.001); women with greater psychopathy traits (total PCL-R score) had lower IQ scores. Negative correlations were also found between IQ and the four PCL-R subscales, Interpersonal, Affective, Lifestyle and Antisocial (r54 = − 0.35, p < 0.01, r54 = − 0.52, p < 0.001, r54 = − 0.53, p < 0.001, and r54 = − 0.49, p < 0.001 respectively).

The results indicate a general negative relationship between PCL-R and IQ, equally distributed across the four subcomponents of the psychopathic trait, and support the view that unsuccessful psychopathic women have poor planning and are unable to foresee and represent future consequences of their actions.

Several studies indicate that cannabis use among patients with psychotic disorders is associated with worse outcome, but only a few studies have controlled for baseline condition and medication.

At 5-year follow-up, interviews were carried out with 314 first-episode psychosis patients included in the OPUS trial. The patients included were in the age range of 18 to 45 years old and 59% were male. Cannabis use was extracted from the Schedule for Clinical Assessment in Neuropsychiatry. At follow-up, the patients were divided into different groups according to the variable cannabis use: abstainers, stoppers, starters and continuers. Psychotic, negative and disorganized dimensions (ranging from zero to five) were calculated for each of the four groups based on the Schedule for the Assessment of Positive and Negative Symptoms in Schizophrenia.

Cannabis users were younger (24.6 years v. 27.4 years, p < 0.001) and had a lower level of education. At the 5-year follow-up, users of cannabis had higher scores on the psychotic dimension [difference 0.97, 95% confidence interval (CI) 0.41–1.53, p = 0.001] and lower levels of the Global Assessment of Functioning (difference 8.26, 95% CI 2.13–14.39, p = 0.01). Those who stopped using cannabis between entry and 5-year follow-up had a significantly lower level of psychotic symptoms at 5-year follow-up even after controlling for baseline level of psychotic symptoms and for insufficient antipsychotic medication (adjusted difference in psychotic dimension –1.04, 95% CI –1.77 to –0.31, p = 0.006).

Continuous cannabis use was associated with higher levels of psychotic symptoms after 5 years, and this association was only partly explained by insufficient antipsychotic medication.

Smoking cessation improves physical health but it has been suggested that in vulnerable individuals it may worsen mental health. This study aimed to identify the short- and longer-term effects of stopping smoking on depression and anxiety in the general population and in those with a history of these disorders.

Sociodemographic and smoking characteristics, and mental and physical health were assessed using established measures in the ATTEMPT cohort, an international longitudinal study of smokers (n = 3645). Smokers who had stopped for at least 3 months or less than 3 months at the 12-month follow-up were compared with current smokers (n = 1640).

At follow-up, 9.7% [95% confidence interval (CI) 8.3–11.2] of smokers had stopped for less than 3 months and 7.5% (95% CI 6.3–8.9) for at least 3 months. Compared with current smokers, prevalence of depression prescriptions obtained in the last 2 weeks was lower for those who had stopped for less than 3 months [odds ratio (OR) 0.37, 95% CI 0.14–0.96] or at least 3 months (OR 0.25, 95% CI 0.06–0.94) after adjusting for baseline prescription levels and confounding variables. Adjusted prevalence of recent depression symptoms was also lower for ex-smokers who had stopped for less than 3 months (OR 0.34, 95% CI 0.15–0.78) or at least 3 months (OR 0.24, 95% CI 0.09–0.67) than among continuing smokers. There was no change in anxiety measures in the general population or any increase in anxiety or depression symptoms in ex-smokers with a past history of these conditions.

Smoking cessation does not appear to be associated with an increase in anxiety or depression and may lead to a reduced incidence of depression.

There is evidence that measures of alcohol consumption, dependence and abuse are valid indicators of qualitatively different subtypes of alcohol involvement yet also fall along a continuum. The present study attempts to resolve the extent to which variations in alcohol involvement reflect a difference in kind versus a difference in degree.

Data were taken from the 2001–2002 National Epidemiologic Survey of Alcohol and Related Conditions. The sample (51% male; 72% white/non-Hispanic) included respondents reporting past 12-month drinking at both waves (wave 1: n = 33644; wave 2: n = 25186). We compared factor mixture models (FMMs), a hybrid of common factor analysis (FA) and latent class analysis (LCA), against FA and LCA models using past 12-month alcohol use disorder (AUD) criteria and five indicators of alcohol consumption reflecting frequency and heaviness of drinking.

Model comparison revealed that the best-fitting model at wave 1 was a one-factor four-class FMM, with classes primarily varying across dependence and consumption indices. The model was replicated using wave 2 data, and validated against AUD and dependence diagnoses. Class stability from waves 1 to 2 was moderate, with greatest agreement for the infrequent drinking class. Within-class associations in the underlying latent factor also revealed modest agreement over time.

There is evidence that alcohol involvement can be considered both categorical and continuous, with responses reduced to four patterns that quantitatively vary along a single dimension. Nosologists may consider hybrid approaches involving groups that vary in pattern of consumption and dependence symptomatology as well as variation of severity within group.

Numerous studies have supported an association between maternal depression and child psychiatric outcomes, but few have controlled for the confounding effects of both maternal and offspring co-morbidity. Thus, it remains unclear whether the correspondence between maternal and offspring depressive and anxiety disorders is better explained by associations between shared features of maternal and offspring internalizing disorders or by specific effects exerted by unique aspects of individual disorders.

Pairs of mothers and offspring overselected for maternal depression (n = 815) were assessed at offspring age 15 years for anxiety and depressive disorders; 705 completed a follow-up at offspring age 20 years. For both mothers and offspring, structural equation modeling was used to distinguish transdiagnostic internalizing pathology – representing the overlap among all depressive and anxiety disorders – from diagnosis-specific forms of pathology. To discriminate between general versus specific pathways of intergenerational transmission of psychopathology, we examined (a) the general association between the maternal and offspring internalizing factors and (b) the correlations between maternal and offspring diagnosis-specific pathology for each disorder.

For mothers and offspring, a unidimensional latent variable model provided the best fit to the correlations among depressive and anxiety disorders. The maternal transdiagnostic internalizing factor strongly predicted the corresponding factor among offspring. In addition, the unique component of post-traumatic stress disorder among offspring was significantly related to the analogous unique component among mothers, but specific components of other maternal disorders, including depression, did not predict corresponding offspring pathology.

Results suggest that intergenerational transmission of internalizing disorders is largely non-specific.

Recent falls in suicide rates should be accompanied by a decline in the prevalence of suicidal ideation.

We used a pseudo-cohort analytic strategy to examine trends in suicidal ideation measured identically in 2000 and 2007, in nationally representative English probability samples of adults aged ⩾ 16 years. Suicidal ideation included tiredness of life, death wishes and thoughts of suicide. Logistic regression models were fitted to estimate trends in age-specific prevalence of suicidal ideation in the past year and past week between 2000 and 2007.

There were 6799 participants aged 16–71 years in 2000, and 6815 participants aged 16–78 years in 2007. There was little evidence of trends in prevalence of suicidal ideation, with the exception of women aged 44–50 years in 2007, whose prevalence was unusually high. Prevalence of suicidal ideation in the past year followed a W-shaped profile with age, with peaks at the transition to adulthood, in the forties, and in the oldest participants.

Despite falling suicide rates, suicidal ideation did not decline overall between 2000 and 2007. This may indicate the success of the National Suicide Prevention Strategy. Women aged 44–50 years in 2007 were, however, particularly prone to suicidal ideation. As they also have the highest age-adjusted prevalence of common mental disorders and the highest female suicide rate, there are clear implications for treatment access, availability and delivery in primary care.

Research into age of onset in obsessive–compulsive disorder (OCD) has indicated significant differences between patients with early and late onset of the disorder. However, multiple criteria have been used arbitrarily for differentiating between early- and late-onset OCD, rendering inconsistent results that are difficult to interpret.

In the current study, admixture analysis was conducted in a sample of 377 OC patients to determine the number of underlying populations of age of onset and associated demographic and clinical characteristics. Various measures of anxiety, depression, co-morbidity, autism, OCD, tics and attention deficit hyperactivity disorder (ADHD) symptoms were administered.

A bimodal age of onset was established and the best-fitting cut-off score between early and late age of onset was 20 years (early age of onset ⩽19 years). Patients with early age of onset were more likely to be single. Early age of onset patients demonstrated higher levels of OCD severity and increased symptoms on all OCD dimensions along with increased ADHD symptoms and higher rates of bipolar disorder.

It is suggested that 20 years is the recommended cut-off age for the determination of early versus late age of onset in OCD. Early age of onset is associated with a generally graver OCD clinical picture and increased ADHD symptoms and bipolar disorder rates, which may be related to greater functional implications of the disorder. We propose that age of onset could be an important marker for the subtyping of OCD.

At present there are no objective, biological markers that can be used to reliably identify individuals with post-traumatic stress disorder (PTSD). This study assessed the diagnostic potential of structural magnetic resonance imaging (sMRI) for identifying trauma-exposed individuals with and without PTSD.

sMRI scans were acquired from 50 survivors of the Sichuan earthquake of 2008 who had developed PTSD, 50 survivors who had not developed PTSD and 40 healthy controls who had not been exposed to the earthquake. Support vector machine (SVM), a multivariate pattern recognition technique, was used to develop an algorithm that distinguished between the three groups at an individual level. The accuracy of the algorithm and its statistical significance were estimated using leave-one-out cross-validation and permutation testing.

When survivors with PTSD were compared against healthy controls, both grey and white matter allowed discrimination with an accuracy of 91% (p < 0.001). When survivors without PTSD were compared against healthy controls, the two groups could be discriminated with accuracies of 76% (p < 0.001) and 85% (p < 0.001) based on grey and white matter, respectively. Finally, when survivors with and without PTSD were compared directly, grey matter allowed discrimination with an accuracy of 67% (p < 0.001); in contrast the two groups could not be distinguished based on white matter.

These results reveal patterns of neuroanatomical alterations that could be used to inform the identification of trauma survivors with and without PTSD at the individual level, and provide preliminary support to the development of SVM as a clinically useful diagnostic aid.

Longitudinal symptoms of post-traumatic stress disorder (PTSD) are often characterized by heterogeneous trajectories, which may have unique pre-, peri- and post-trauma risk and protective factors. To date, however, no study has evaluated the nature and determinants of predominant trajectories of PTSD symptoms in World Trade Center (WTC) responders.

A total of 10835 WTC responders, including 4035 professional police responders and 6800 non-traditional responders (e.g. construction workers) who participated in the WTC Health Program (WTC-HP), were evaluated an average of 3, 6 and 8 years after the WTC attacks.

Among police responders, longitudinal PTSD symptoms were best characterized by four classes, with the majority (77.8%) in a resistant/resilient trajectory and the remainder exhibiting chronic (5.3%), recovering (8.4%) or delayed-onset (8.5%) symptom trajectories. Among non-traditional responders, a six-class solution was optimal, with fewer responders in a resistant/resilient trajectory (58.0%) and the remainder exhibiting recovering (12.3%), severe chronic (9.5%), subsyndromal increasing (7.3%), delayed-onset (6.7%) and moderate chronic (6.2%) trajectories. Prior psychiatric history, Hispanic ethnicity, severity of WTC exposure and WTC-related medical conditions were most strongly associated with symptomatic trajectories of PTSD symptoms in both groups of responders, whereas greater education and family and work support while working at the WTC site were protective against several of these trajectories.

Trajectories of PTSD symptoms in WTC responders are heterogeneous and associated uniquely with pre-, peri- and post-trauma risk and protective factors. Police responders were more likely than non-traditional responders to exhibit a resistant/resilient trajectory. These results underscore the importance of prevention, screening and treatment efforts that target high-risk disaster responders, particularly those with prior psychiatric history, high levels of trauma exposure and work-related medical morbidities.