Introduction

On April 6, 2009, a 6.3-magnitude earthquake hit L’Aquila, a town in central Italy with 70,000 inhabitants. In 30 s, the earthquake caused 309 victims and 1,500 injured, and largely destroyed the town’s historic center as well as thousands of houses [1].

Evidence shows that there are specific personal and social behavioral changes after wars or natural catastrophes [2–4]. When an individual is exposed to a very stressful event, the behavioral response is hyperactivation or paralysis and the somatic response is characterized by hyperactivity of the autonomous nervous system and the hypothalamus–pituitary–adrenal axis (HPA) [5, 6].

Post-traumatic syndrome is represented by a continuous state of alarm, even in the absence of direct negative stressors, and the post-traumatic stress disorder (PTSD) is a complex psychopathological disease that could develop after a traumatic event [7, 8].

Its assessment and treatment are problematic, as its symptoms belong to the anxious/depressive spectrum, with somatic correlates [6, 9]. The principal symptoms of PTSD are classified into three clusters: intrusion, avoidance/numbing and hyperarousal [6, 8]. However, other investigations have discussed about the criteria and the prevalence of PTSD, on the basis of the use of Diagnostic and Statistical Manual of Mental Disorders fourth edition (DSM-IV) or DSM-V [10, 11].

After L’Aquila earthquake, many empirical studies on effects of trauma on particular local populations regarding the adolescents, the gender differences, the pregnant women and the psychotic patients were made [3, 12–14].

Regarding the prevalence, Dell’Osso et al. [15] have found a percentage of 37.5 % of PTSD cases in a large sample of students, assessed with self-report psychometric tools. On the other hand, other researchers, using telephonic interviews, found only the 4.1 % of people affected with post-traumatic stress [16]. Generally, the great variability of data regarding the prevalence of PTSD after collective traumas depends on the methodology of investigation and on the severity of the disaster, with a range of prevalence between 4 and 87 % [17].

Moreover, the personal vulnerability to PTSD is caused by many endogenous and exogenous factors. In this regard, the coping strategies play a fundamental role against PTSD.

Lazarus [18] classified coping responses as either problem-focused or emotion-focused. In a study concerning coping strategies in the mental diseases, Meyer made a further subdivision in adaptive and maladaptive coping, on the basis of Carver’s model [19].

Another very significant aspect regards the physical illnesses in the population after catastrophes or collective traumas. PTSD, with its psychobiological consequences, also has a role in chronic and metabolic pathologies, as diabetes [20]. In fact, some studies also suggest that chronic or acute distress may affect the metabolism of glucose and increase the risk of developing type 2 diabetes [21–26]. In particular, an important and recent investigation demonstrated the link between PTSD and diabetes in a large study population [27], showing a possible cause–effect link between traumatic stress and diabetes due to the dysregulation of HPA and the related increase of corticotropin and cortisol, that plays a role in the pathogenesis of diabetes [28]. On the whole, it is known that many psychological elements, as depression, anxiety and mood disorders, are involved in the diabetes [29–32].

Given this background, the primary hypothesis of this study concerns the people that survived the earthquake and after were diagnosed with diabetes. Therefore, we hypothesized a major prevalence and severity of PTSD in patients diagnosed with diabetes after the earthquake.

Moreover, we know that PTSD occurs more in women that in men, and this evidence verified also in the local general population. In addition, literature affirms that in diabetic people another mental disease, as depression, specifies female diabetic population. Considering these aspects, we also hypothesize a possible major prevalence and a higher severity of PTSD in women with diabetes than in men. Hence, the secondary hypothesis regards the gender differences with a major severity of PTSD in women with diabetes than in men [3, 33–35].

Objectives

According to our hypotheses, the first aim of this investigation focuses on the psychometric evaluation of PTSD and the coping strategies in type 2 diabetic patients diagnosed at least 6 months after the L’Aquila earthquake of April 6, 2009.

The second aim regards the possible gender differences in PTSD and in the use of coping strategies in our selected diabetic patients.

Materials and methods

Sample recruitment

At the Unit of Diabetology and Metabolic Diseases of San Salvatore Hospital, L’Aquila, we consecutively recruited 100 patients aged 30–65 with diabetes mellitus type 2 who had lived through the earthquake. This sample represented about the 10 % of the local diabetic population, considered on the basis to our inclusion criteria. Sixty of these had suffered from diabetes before the earthquake (pre-quake patients). The remaining 40 were diagnosed at least 6 months after the earthquake (post-quake patients). Into the entire group, we enrolled 54 men and 46 women.

The inclusion criteria were to be resident in L’Aquila, to have lived through the earthquake and to have lived post-quake in a hotel, tent, camper or caravan for between 1 and 6 months.

We excluded all patients with vascular, neurological, renal, retinal or hormonal complications and with diabetes diagnosed more than 10 years earlier. We also excluded patients who had suffered major events such as death in the family caused by the earthquake or who were physically injured by the collapse of their home, or who had a personal story of previous traumatic experiences.

A clinical psychologist assessed all subjects according to DSM-IV, through a clinical interview, to exclude severe mental disorders, as psychosis or major depression, while a diabetologist examined the patients for blood glucose levels and general health.

Hence, a first part of our study regarded the following subdivision: pre-quake patients and post-quake patients. A second part regarded the gender differences into diabetic people.

All assessments and data refer to the 6-month period from June to December 2012.

The study protocol was approved by the San Salvatore Hospital ethics committee for investigations involving human subjects in line with the Declaration of Helsinki, and all patients signed their informed consent to the handling of personal data.

Assessment

Two self-report tests were used: the Davidson Trauma Scale (DTS) [36] and the Brief-COPE [19, 37]. The Davidson Trauma Scale is a self-report questionnaire composed of 17 items with four possible Likert scale responses exploring both severity and frequency of symptoms. All 17 items explore the first open question of the test (what was the most traumatic event for you?). This psychometric tool has been validated for PTSD diagnosis relative to a specific traumatic event and excluding other possible traumas. The cutoff total score for PTSD diagnosis is ≥40. The correction grid indicates the partial scores for PTSD severity and frequency as well as the severity of the three main symptoms of post-traumatic stress: intrusion, avoidance/numbing and hyperarousal.

The Brief-COPE is a well-validated psychometric tool structured in a 28-item test with a 4-response Likert scale evaluating 14 specific coping strategies. By using Brief-COPE, we evaluated two coping domains: adaptive coping (the first eight scales) and maladaptive coping (the last six scales) [38].

For the biological assessments, we collected blood sticks picked up in the morning between 8 and 10, before breakfast. These were tested in the diabetes center according to the normal clinical workup. All biometric assessments were collected once, in the same period as the administration of the psychometric tools.

Statistical analysis

The power of analysis indicated that with a local diabetic population, composed of about 4,000 patients and an estimated prevalence of PTSD of 25 %, a sample of 100 diabetic patients would provide a 95 % confidence interval representing 10 % of the considered population.

To compare continuous variables, Student’s T test for non-matched data and assuming an equal variance was used.

The difference between categorical variables was tested using Chi-square test or Fisher’s exact test, when appropriate. Chi-square test was chosen to investigate the differences in the prevalence of PTSD in the groups. Univariate logistic regression was employed to verify possible predictive factors at time of PTSD, assuming DTS cutoff score (presence/absence of PTSD) as dependent variable and coping strategies, years of diabetes and age as independent variables. The statistical analysis was made with MedCalc software.

Results

PTSD in newly diagnosed patients

In the first 100 patients attending our diabetic clinic and fulfilling our inclusion criteria, 60 patients had been diagnosed with diabetes before the earthquake and 40 after it. Table 1 shows their sociodemographic characteristics. The mean age of the first group was 56.3 ± 8.38 years, comparable to that of the second group (54.9 ± 10.02 years). There were no significant differences between these two groups in education and in the prevalence of gender. There were no differences in the use of insulin between the two diabetic patient groups.

Table 1 Sociodemographic and clinical characteristics according to the epoch of the diagnosis
Full size table

The DTS analysis showed a difference among the two groups for the indication of the “most traumatic event.” In fact, 80 % of subjects in the post-quake group but only 30 % in the pre-quake and control group indicated the earthquake. This difference was statistically significant (p < 0.0001; χ 2 = 21.78). It is interesting that only 5 % of post-quake patients indicated diabetes as the most traumatic event, in comparison with 25 % of pre-quake patients, also in this case there was a statistically significant difference (p = 0.0002; χ 2 = 14.157). Moreover, in the post-quake group, other forms of traumatic events were less important than the earthquake, while other traumatic events were largely reported in the pre-quake (45 %) (Table 1).

Using the DTS cutoff score for PTSD, we found that 67.5 % of patients in the post-quake group had PTSD, in comparison with 33.3 % in the pre-quake group. Here too, there was a statistically significant difference between the post-quake group and the pre-quake group (p < 0.0001; χ 2 = 21.78) (Table 1).

PTSD was not only more prevalent but also more severe in post-quake patients than in the other group. As shown in Table 3, the mean total DTS score was 51.72 ± 26.05 in the post-quake group, 31.65 ± 22.59 in the pre-quake group, revealing a significant difference (p = 0.0001; t = 4.092).

Regarding the DTS sub-domains, there were statistically significant differences in avoidance/numbing and hyperarousal between the post-quake and pre-quake groups (Table 3).

The analysis of results for the Brief-COPE questionnaire revealed an interesting statistical divergence for maladaptive coping. In fact, the mean score was significantly higher in the post-quake group (26.17 ± 4.64) than in the pre-quake group (22.37 ± 5.49) (p = 0.0005; t = 3.602) (Table 3).

There was no significant difference in the biological parameters of the two groups of diabetic patients (Table 3).

Finally, Table 5 shows that maladaptive coping was the specific and only predictive psychological factor for PTSD in the post-quake group only (p = 0.006; OR 1.682).

Gender differences

Table 2 lists the mean age of males and female groups was 56.4 (SD = 5.6) and 54 (SD = 11.5) years, respectively, with no significant difference, such as in the level of education (Table 2).

Table 2 Sociodemographic and clinical characteristics according to the gender
Full size table

As regards a cutoff score of 40 to establish a PTSD condition, a significant difference (p = 0.0001; χ 2 = 15.872) was observed between men and women: 29.6 % (16/54) and 58.7 % (27/46), respectively (Table 3).

Table 3 Psychological and biological aspects in pre-quake and in post-quake patients
Full size table

When we observed the most traumatic event in the two groups, 33.3 % (18/54) in the male group and 51.1 % (24/46) in the female group indicated the earthquake. This comparison indicated a significant difference (p = 0.01; χ 2 = 5.932); on the other hand, 38.8 % (21/54) of males and 8.7 % (4/46) of females indicate diabetes as the most traumatic experience. Also in this case, there was a remarkable difference between the two groups (p < 0.0001; χ 2 = 23.054).

Another statistical difference was found in the means of DTS total score: 31.65 (SD = 23.06) in males and 53.5 (27.01) in females (p < 0.0001; t = 4.364). Moreover, we found differences between males and females in symptoms of PTSD such as intrusion, avoidance/numbing and hyperarousal (Table 4).

Table 4 Psychological and biological aspects in male and female diabetic patients
Full size table

Biological parameters are compared in Table 4, but the only statistically significant difference was in the values of blood pressure (Table 5).

Table 5 Univariate logistic regression among coping strategies, years of diabetes and age on PTSD in pre- and post-quake patients
Full size table

Univariate logistic regression showed that only in males, coping strategies had an influence on PTSD, both adaptive and maladaptive coping: the first one in protective way (p = 0.01; OR 0.87) and the second one with a predictive aspect respect to PTSD (p = 0.007; OR 1.25) (Table 6).

Table 6 Univariate logistic regression among coping strategies, years of diabetes and age on PTSD in women and men
Full size table

Discussion

According to our primary hypothesis, these results demonstrate that there is a strong link between diabetes and PTSD in patients diagnosed with diabetes after the earthquake. For these patients, the natural disaster was the most traumatic event in their life, and the severity of their post-traumatic symptoms was higher than in pre-quake patients. In fact, we observed that the earthquake was not a significant psycho-traumatic event for the patients with pre-quake diabetes.

In patients diagnosed before April 6, 2009, post-traumatic syndrome has been of minor importance and in line with other studies on local population [15, 39] or after other natural catastrophes [17, 40]. In other words, only the newly diagnosed diabetic patients (post-quake group) were seriously affected by PTSD after the earthquake. Hence, the most striking result was undoubtedly the high prevalence and greater severity of PTSD in people who developed diabetes after the earthquake.

We could suppose that the recent diagnosis of diabetes had affected the mental health of these patients, but regression analysis related to the impact of duration of diabetes on PTSD and the capacity to discriminate of Davidson Trauma Scale did not confirm this hypothesis [36].

It cannot therefore be excluded that in some patients the exacerbation of PTSD after the earthquake had a role in the earlier development of diabetes. In this case, PTSD could be considered as one of the risk factors for diabetes and also a comorbidity factor.

In fact, different studies state that chronic stress and PTSD could be predictive of a metabolic disorder such as diabetes. [21, 27, 41–43]. The possible diabetogenic effect of post-traumatic stress is due to the dysregulation of cortisol or catecholamines during the hyperactivation of the HPA and the autonomous nervous system [6, 21, 44, 45].

Nevertheless, we did not found differences in our two groups of patients in the indicators of diabetes and its control: blood glucose and HbA1c [24].

Moreover, our findings revealed for the first time another important aspect. After a traumatic event, diabetic women develop PTSD more often than men.

It is known that in the general population and also in the local population, PTSD is more present in women than in the men [2, 3], but in a particular clinical population such as diabetic patients, this evidence was still unknown.

Women who were the subject of our investigation showed higher PTSD levels. It is likely that women had more negative thoughts with respect to stressful events with greater implications or traumatic memory [46].

In this case, in our female group, PTSD adds to the severe condition of chronic pathologies such as diabetes. Women were more affected by avoidance and numbing to demonstrate the pervasive effect of trauma, but they also revealed higher levels of hyperarousal symptoms that could be explained through the various somatic correlates of post-traumatic stress [9]. In this dramatic and psychopathological situation, we may suppose that women show many problems related to glycaemic control, but our results did not support this aspect [47]. In fact, diabetic females did not differ from males with regards to diabetes indicators (BMI, glycaemia and glycated hemoglobin). Nevertheless, we can validate our second hypothesis regarding gender differences in PTSD in diabetic people.

Another important aspect to strengthen the higher prevalence and severity of PTSD in diabetic women regards the causes of trauma. Women reported the earthquake as the most traumatic event, whereas men reported diabetes. Men give less importance to natural catastrophes; on the other hand, they consider chronic pathologies more dangerous and disabling. It may seem that women are more focused on negative collective events, whereas men are more concerned with negative personal health problems with a possible comorbidity of adjustment disorder.

These differences that we found in the severity and prevalence of PTSD in patient newly diagnosed and in women patients could be explained by personal characteristics, such as use of coping strategies [48–50]. In fact, our data revealed that in the patients diagnosed post-quake, maladaptive coping played a fundamental role and it was partially predictive of the post-traumatic syndrome. In the pre-quake groups, where PTSD was less evident, maladaptive coping had no significant impact on post-traumatic stress. In this group, we could suppose that the coping skills are more efficacies to front traumatic stress for a quickly reason: These patients received from more time the clinical care and the medical support by physicians and nurses for their metabolic disease. The first crisis due to diabetes diagnosis was adequately supported by care and the subsequent earthquake, and the possible exacerbation of PTSD was few impacting compared to post-quake patients. Hence, from this point of view, the medical care for diabetes could be considered as a protective factor against other life stress event and the physician–patient relationship could play an important role for an improvement of the coping strategies.

On the contrary, the post-quake patients were more vulnerable to PTSD for the negative effect of maladaptive coping, and for this reason, it is necessary to implement the care in these subjects also for the psychological aspects.

Then, we made an analysis regarding the use of coping strategies in men and women; we noticed that only in male group there was a double action of coping strategies, at the same time in protective and predictive terms. In males, there was a possible attempt of response to trauma, with the recourse to positive and negative mechanisms of reaction linked to PTSD. In women, where PTSD is more pervasive, it is likely that are activated responses of seclusion that characterize the depressive feeling due to a collective trauma.

Limitations

The first limit is represented by the low number of sample subjects concerning a specific local population, and an extension of these data on other population after natural disaster is necessary.

The second concerns the observational design of the study that obviously omits the clinical history especially for the diabetic patients newly diagnosed. In this regard, it is lacking the assessment of past psychological symptoms, as anxiety or mood disorders such as information about hormonal patterns related to the physiology of the HPA axis. An accurate assessment of these aspects in longitudinal way will be an object of future investigations.

Conclusions

Our results revealed greater levels of PTSD in post-quake patients than in pre-quake confirming the primary hypothesis of this study about the complex psychoneuroendocrinological link among stressful life events, trauma and diabetes.

Moreover, these evidences have highlighted, as expected and according to our secondary hypothesis higher levels and a major prevalence of PTSD in women compared to men.

Therefore, we suggest that PTSD may be considered as an important comorbidity factor in newly diagnosed patients and in diabetic women after a traumatic event. On the whole, in new diabetic patients and particularly in women, psychological support seems to be very important to help them react to both PTSD and diabetes and to help them improve their coping skills.