Association of Acyl-Ghrelin With Posttraumatic Stress Disorder in Adolescents Who Experienced Severe Trauma

This cross-sectional study examines acyl-ghrelin levels in Pakistani adolescents who experienced severe trauma and have posttraumatic stress disorder (PTSD) compared with adolescents who have not experienced severe trauma.

from significant illness (Dengue Fever, etc.) or who currently showed signs of acute or chronic illness by physical examination. Thirteen children were excluded from the study for these reasons; no data was included from these subjects in any analysis.

Salivary Sample Collection
Some subjects (n = 37 traumatized, n = 23 control) consented to the collection of an early morning saliva sample to measure cortisol (at 8 A.M.). Some subjects may not have chosen to participate in this portion of the study because of the early time, or because it extended the time commitment required such that the study visit spanned a full day. These participants were asked not to exercise or eat for at least 30 min prior to saliva collection. At the time of sample collection, subjects were instructed to place a cotton ball in their mouth, roll it with their tongue, and gently chew for 1 min to saturate the cotton ball. The saturated cotton ball was squeezed into a sterile microfuge tube. The tube was rapidly transferred to a -80 o C freezer until analysis. The analysis of each saliva sample was performed within one month of collection.

Demographic and Clinical Data Collection
For all subjects (n = 49 traumatized, n = 39 control), a structured, face-to-face interview was performed on the afternoon of the study visit. Subjects were instructed to not eat after lunch so that a fasting venous blood sample could be collected. Children and their caregivers were both present throughout the visit. Questionnaires were administered in Urdu and were translated to English. All subjects received venipuncture to collect a blood sample for the analysis of acylghrelin (at 5 P.M). During a clinical evaluation, occurring immediately before or after venipuncture, all subjects were administered the PTSD CheckList-Civilian Version (PCL-17; hereafter referred to as PCL) translated into Urdu by four experimenters fluent in both Urdu and English (see Supplement, Questionnaire). This questionnaire has a high correlation with the results of the Clinician-Administered PTSD Scale-Child/Adolescent Version (CAPS-CA) 2 .
Anthropomorphic and health information for each child was collected during the interview. The questions addressed changes in health status since the time of the terror attack that resulted in the death of a loved one or injury to the child (for children in the Traumatized groups) and changes in the health status since the time of the natural death of a loved one (if applicable) or in recent years (for children in the Control group). Additionally, parents were asked to provide information about their education levels, occupations, and household income (in Pakistani rupees).

Venous Blood Sample Collection
The protease inhibitor 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride (AEBSF; 200 mg/ml) was mixed in distilled water and maintained on ice prior to sample collection. A small volume of blood (3 ml) was drawn by trained research staff. Within two minutes of blood collection, aliquots of blood (100 µl) were generated; AEBSF was added to each aliquot at a 1:100 dilution (i.e. 1 µl AEBSF per 100 µl of blood). Samples were gently inverted five times to mix completely, and then placed on ice. Serum was then collected from the clotted blood aliquots and stored at -80 o C until analysis.

Processing of Saliva and Serum Samples
For salivary samples, a salivary cortisol ELISA kit (CO116S; Calbiotech Inc., 10461 Austin Dr, Spring Valley, CA) was used. Acyl-ghrelin was measured using an ELISA kit as previously described 1 . For all ELISAs, the results were analyzed by experimenters not involved in the sample collection who were blinded to the group information associated with the sample.
Samples were run in duplicate and mean values were used in subsequent analyses. All reported values were dilution-corrected.

PCL Scoring
For the PCL, a total score was created by summing the responses across all 17 items. Responses to individual items were classified as symptomatic if rated at "3: Moderately" or above. Subjects were classified as having PTSD (PTSD+) if they were symptomatic for one or more Reexperiencing items (items 1-5), three or more Avoidance/Numbing items (items 6-12), and two or more Arousal items (items 13-17). All subjects that were not classified as PTSD+ were classified as not having PTSD (PTSD-).

Socioeconomic Status Scoring
To calculate socioeconomic status, the Kuppuswamy classification scale was used 3

. The
Kuppuswamy classification scale is the most commonly used measure of socioeconomic status in Pakistan and is updated regularly to index the income scales to changes in the consumer price index for Pakistan. This scale assigns a numerical score to the educational level attained by the heads of the household, the occupations of the heads of the household, and the household income. These three scores were summed to generate a total score. The total scores were converted to class ratings (upper class: 26-29; middle class: 11-25; lower class: 3-10). These ratings were recoded as categorical variables (upper class = 3, middle class = 2, lower class = 1).

Body Mass Index [BMI] was computed by dividing body mass (kg) by height squared (m 2 ).
Statistical outliers for each measure were identified as values that were more than three standard deviations from the group mean. One data point was excluded from analysis of the Morning Cortisol for this reason (from the Traumatized, PTSD+ Group); no other outliers were identified.
Normality of data was analyzed with the D'Agostino & Pearson omnibus normality test. Group means of continuous variables from a normal distribution were compared using one-way ANOVA followed by Tukey's multiple comparisons. Group means of continuous variables that were not normally distributed were compared using either the Kruskal-Wallis test (for three- were not correlated (low multicollinearity). For these linear regressions, acyl-ghrelin, morning cortisol, and age values were log-transformed (base 10) to meet normality requirements for the regression.
To quantify the elevated risk of PTSD associated with elevated acyl-ghrelin, we performed an odds ratio (OR) analysis. Each traumatized subject was classified as having low, moderate, or high levels of acyl-ghrelin. The distribution of acyl-ghrelin values in the traumatized population was divided into thirds, with an equal number of subjects in each third. The lowest third of values were classified as "low, the middle third as "moderate", and the upper third as "high". A logistic regression was run to determine the association of the low and moderate categorical acylghrelin values with the binary outcome of PTSD diagnosis (Yes/No). An odds analysis for subjects with high acyl-ghrelin levels was not performed because all subjects with high ghrelin levels had PTSD (the odds ratio for this group is infinity).

Questionnaire
The PCL-17-C was translated into Urdu and is provided below. The English version of the PCL-17-C was developed by the Veterans Affairs National Center for PTSD and is also provided below.