A Chinese Herbal Formula to Improve General Psychological Status in Posttraumatic Stress Disorder: A Randomized Placebo-Controlled Trial on Sichuan Earthquake Survivors

2.1. Study Design and Setting

Patients were enrolled into this study five months after the 2008 Sichuan earthquake, between October 2008 and January 2009, through a community-based epidemiological survey of four settlements of a severely affected city, Dujiangyan. In the enrollment survey, the relationship between exposure to the earthquake and PTSD was assessed. Preliminary screening was performed in the communities by our researchers according to the DSM III for PTSD, Chinese version [40]. Eligible subjects were invited to participate in a diagnostic face-to-face or telephone interview with one of three experienced psychiatrists, each of which has at least eight years of clinical experience. Patients who met the inclusion and exclusion criteria were enrolled (see Patient Flow Chart, Figure 1), and our psychologists verified PTSD as the primary diagnosis of each enrollee. Inclusion criteria were age 16 or older, meeting DSM III criteria for PTSD with at least one of the nine Symptom Check-List-90-Revised (SCL-90-R) [41] subscores above the Chinese norm [42], and being willing to be randomly assigned. Participants understood that those randomized into the placebo control group could receive XTJYF after completion of the whole trial if they wished. Exclusion criteria were past history of bipolarism, schizophrenia, or other psychotic disorders; current organic mental disorder, factitious disorder, or malingering; any past history of alcohol or substance dependence or abuse; evidence of clinically significant hepatic or renal disease or any other acute or unstable medical condition that might interfere with safe participation in the study; use of any medication with clinically significant psychotropic activity within two weeks of randomization; any cognitive-behavioral therapy during the trial; psychotherapy initiated or ending during the trial. For female patients of childbearing age, participation was contingent on a negative serum pregnancy test and a medically accepted method of contraception.

Written informed consent was obtained from all patients before participation. Patients were free to withdraw from the study at any time. Clinical diagnoses, physicals, and laboratory examinations were mainly conducted in the outpatient clinic at the Air Force Sanatorium in the city of Dujiangyan by our psychologist and other investigators. The research staff collected patients' weekly feedback on their medical conditions and delivered the XTJYF or placebo through in-house visits. The trial protocol was approved by the Ethics Committee of Shanghai Changzheng Hospital and the Air Force Sanatorium in Dujiangyan.

A sociodemographic inventory and a medical history were taken, and a routine physical and laboratory examination (i.e., blood pressure, ECG, clinical chemistry and hematology tests, and urinalysis) was performed by the investigators as a baseline for future toxicology screening.

2.2. Randomization and Blinding

Eligible patients were randomized to either XTJYF treatment or placebo control. Random numbers were generated by computer software; treatment codes were held by the chief investigator, who was isolated from patients and outcome data. The chief investigator was also responsible for distributing the XTJYF and placebo with the assistance of our research staff. Patients, research staff, and data entry clerks were blinded to treatment group assignment. Treatment compliance was assessed by package count and observation by the research staff. Treatment codes were disclosed after the entire study was completed.

2.3. Study Interventions

All patients received 12 g packages of granulated XTJYF or placebo twice a day for eight weeks [39] and were instructed to drink the contents dissolved in warm, boiled water.

2.4. Outcome Measures

Each patient completed the SCL-90-R questionnaires twice, at baseline prior to randomization and in the eighth week after the randomization, that is, at the end of this clinical trial. The SCL-90-R is a questionnaire for self-reporting psychological distress. It is widely used in patients suffering from mental diseases and for psychological evaluation of healthy individuals. The instrument is well accepted for its good internal consistency, dimensional structure, reliability, and validity [43–45]. The Chinese SCL-90, translated and validated by Wang from the English version of the SCL-90-R, was used [46, 47].

The SCL-90-R consists of 90 symptoms of distress. Patients were instructed to indicate the degree to which they had been troubled by each symptom during the preceding week by ranking the symptom from 0 to 4, with 0 being “not at all” and 4 being “extremely.” The statements were classified into nine dimensions, or factors (F), that reflect various types of psychopathology: (F1) somatization, (F2) obsessive-compulsive behavior, (F3) interpersonal sensitivity, (F4) depression, (F5) anxiety, (F6) hostility, (F7) phobic anxiety, (F8) paranoid ideation, and (F9) psychoticism. Three supplementary global indices reflect the degree of symptomatology. The global severity index (GSI) registers the average depth of impairment based on the severity recorded for each symptom; the positive symptom total index (PST) indicates the total number of symptoms experienced; the Positive Symptom Distress Index (PSDI) reflects the level of distress by correlating the reported symptoms [41]. In addition, on the SCL-90-R, there are seven items not included in any of the nine factors, among which, three reflect sleep quality. Individual SCL-90-R factors have been used to evaluate the psychological condition of PTSD patients, and there is sufficient evidence to support the correlation of higher global SCL-90-R scores with the severity of a patient's core PTSD symptoms [12, 48–57].

During the trial, patients were closely monitored for adverse events (AEs) and worsening of symptoms. The time of onset of any observed or spontaneously reported AE, its duration and severity, any action taken, and the outcome were recorded.

2.5. Herbal Preparation and Dispensing

The original formula, Xiao-Yao-San, contains eight herbs: Chai-Hu (Radix Bupleuri), Dang-Gui (Radix Angelicae sinensis), Fu-Ling (Poria), Bai-Zhu (Rhizoma Atractylodis macrocephalae), Bai-Shao (Radix Paeoniae alba), Bo-He (Herba Menthae), Zhi Gan-Cao (Radix Glycyrrhizae preparata), and Sheng-Jiang (Rhizoma Zingiberis recens). Our modification, XTJYF, contains all the herbs of the original formula, except Sheng-Jiang, plus additional seven herbs, including Fa Ban-Xia (Rhizoma Pinelliae preparatae) and Chen-Pi (Pericarpium Citri reticulatae), that are commonly used for psychological disorders (see Table 1).

All herbal substances used in this trial are listed with the Pharmacopoeia Commission of China, 2005, and are accepted as suitable for human consumption when administered within standard dosage levels. None of these herbs is a controlled substance or an endangered species. Raw herbs were purchased from the Lei Yun Shang Pharmaceutical Company (Shanghai, China). The herbs were extracted with water, and the resulting granules were packaged by the Chinese Drug Preparation Department of Shanghai Changzheng Hospital. Levels of heavy metals and microbial and pesticide residues were carefully assessed, and all fell well within the normal range [58].

The placebo granules, purchased from Jiangsu Tianjiang Pharmaceutical Company, Ltd., were designed to resemble the XTJYF granules in taste, smell, and appearance. The placebo was composed of dextrin, sunset yellow fcf, and a sweetener; the proportion was 1200 : 1 : 7. After being tested on five independent volunteers, the placebo was deemed indistinguishable from XTJYF. XTJYF and the placebo were dispensed in identical opaque packages.

2.6. Statistical Analysis

Quantitative data was summarized using mean, standard deviation (SD), or 95% confidence interval (95% CI). Qualitative data was described using proportion, as percentages. Baseline characteristics of the two groups were compared using the two-sided chi-square test or t-test at a significance level of 0.05.

Since this was a randomized, blind clinical trial, the statistical analyses for treatment effect evaluation of the primary and secondary outcomes are relatively straightforward. Baseline-to-end-point score changes in the three global SCL-90-R indices and rates of response in the GSI were computed as the primary endpoints. For defining rate of response, patients with a reduction of at least 30% from the baseline GSI score were classified as “much improved”; at least 50%, as “very much improved.” Subanalyses of the baseline-to-end-point score changes of the nine SCL factors and sleep quality score (the average of the scores of the three SCl-90-R items on sleep quality) were secondary endpoints. Statistical analysis on both primary and secondary outcomes was done using intention-to-treat analysis (ITT) with statistical software SPSS. Missing values in the SCL-90-R questionnaire for the patients who withdrew from the study before the eighth week were imputed using the last-observation-carried-forward method. For primary outcomes, effect sizes (for three global indices) and number needed to treat (NNT, for rate of response in the GSI), as well as the P values from two sample t-tests and chi-square tests, are reported in the treatment effect assessment. The same analytic approaches were applied to the secondary outcomes. Additionally, Fisher's exact test was used to compare the difference in dropout rate and AEs between the two treatment groups.

3.1. Baseline Characteristics and GPS Assessment

Table 2 shows that randomization was effective and that there were no significant differences between the two groups in baseline demographics, core clinical PTSD symptoms, or baseline SCL-90-R global indices. Even though individual SCL-90-R factor scores and sleep quality scores at baseline are not shown here, we checked all of them and founded no significant differences between the two groups. Notice that women constituted 72% of XTJYF-treated and 71% of placebo-treated patients. Ages ranged from 16 to 85; 64% were over 45.

Table 3 shows the urgency of the public health needs of these earthquake-affected PTSD patients and indicates that baseline scores of the patients in our clinical trial are significantly higher in all nine SCL-90-R factors and all three supplementary global indices than those seen in Chinese and American norms [41, 42].

3.2. Treatment Effect on Primary Outcomes

Table 4 shows that patients in the XTJYF group experienced statistically significant improvement after treatment in all three supplementary global index scores compared to the placebo group. Based on the reported effect sizes, XTJYF treatment has a moderate effect on GSI and PSDI indices and a small effect on the PST index. Our findings on the rate of response, defined based on GSI score improvement, are displayed in Figure 2; 50% of the XTJYF patients versus 28% of those in the placebo group were “much improved,” providing statistically significant evidence supporting the advantage of XTJYF over placebo at the level of 0.05 (P value = 0.001). The NNT is 4.55. Additionally, as Figure 2 shows, 20% of the XTJYF patients versus 12% of those in the placebo group were “very much improved,” but this result is not statistically significant (P value = 0.12).

3.3. Treatment Effect on Secondary Outcomes

The second part of Table 4 displays the treatment effects of XTJYF and placebo on the nine SCL factors and sleep quality score. The results indicate that, in comparison to placebo, the XTJYF group experienced statistically significant improvement after treatment in five of the nine SCL factors, somatization (P = 0.003), obsessive-compulsive behavior (P = 0.036), depression (P = 0.001), anxiety (P < 0.001), and hostility (P = 0.019). Based on the reported effect sizes, XTJYF treatment has a moderate effect on somatization, depression, anxiety, and hostility, as well as a small effect on obsessive-compulsive behavior, interpersonal sensitivity, and phobic anxiety. Table 4 also shows that XTJYF treatment yielded statistically significant improvement in sleep quality at the end of the study, with a P value of less than 0.001 and a moderate effect size.

3.4. Treatment Tolerability

Overall, XTJYF was well tolerated. Compliance rate, 83% for the XTJYF and 81% for the placebo group, was reasonably high. Six in the XTJYF and five in the control group withdrew due to adverse effects, so reported AEs were similar in the two groups. The most frequently reported AEs were nausea (14.6% versus 9.0%; P = 0.24), diarrhea (10.6% versus 6.5%; P = 0.36), and malaise (10.6% versus 12.3%; P = 0.69). All AEs were minor and were determined to be unrelated to the ingestion of XTJYF. In the XTJYF group, 21 subjects dropped out (17.1%); in the placebo group, 23 did (18.9%, P = 0.74). The primary reasons cited for dropout in the XTJYF and placebo groups, respectively, were AE (4.9% versus 4.1%; P = 1); lost to follow-up (1.6% versus 2.5%; P = 0.68); protocol violation (4.9% versus 3.3%; P = 0.75); lack of efficacy (3.3% versus 6.6%; P = 0.25); miscellaneous reasons, for example, disliked the taste of the herbs (2.4% versus 2.5%; P = 1). Subjects' laboratory values and vital signs were similar in the two groups. Changes in these values were minor, infrequent, and not considered clinically meaningful.