Psychological intervention and its immune effect in cancer patients: A meta-analysis.

OBJECTIVE: To determine whether psychological intervention (PI) changes the levels of immune indicators in cancer patients.
METHODS: We conducted a systematic search published up to July 2018, followed by a manual search. Randomized controlled trials were included. Two reviewers independently screened and extracted data, which were analyzed using Review manager 5.3.
RESULTS: Twenty-nine studies were included including four kinds of PI. Only stress management didn't result in immune changes; only cognitive behavior therapy affect NK cell activity. PI did not change immune indicators on cancer patients who completed therapy. Compared to patients not receiving PI, those received PI had significantly higher NK cell count and activity in whole blood; and serum levels of IL-2, IL-4, IFN-γ, lgA, and lgG. However, the differences in the serum levels of IL-6, IL-10, TNF-α, and IgM were not significant (P > .05), and the changes recorded for the CD3, CD4, and CD8 cell count, and CD4/CD8 ratios were inconsistent.
CONCLUSIONS: Although there are considerable evidences of PI's immune effect, but its magnitude was moderate. Therefore, it may be premature to conclude whether PI affects immunity of cancer patients. Further research is warranted, with special focus on the PI types and treatment methods.

Introduction

Cancer is an important public health concern worldwide. GLOBOCAN 2012 reported that there were 14.1 million new cancer cases, 8.2 million cancer deaths, and 32.6 million people living with cancer (within 5 years of diagnosis) in 2012 worldwide.[ Traditional cancer treatments, such as surgery, chemotherapy, and radiotherapy, certainly affect the medical outcomes of cancer, but may not completely eradiate all types of cancers and always cause adverse effects. Therefore, enormous efforts are invested in exploring adjunctive interventions with minimal adverse effects in cancer patients.[

Etiological studies have shown that genetic, environmental, and socioeconomic factors are only partly responsible for the development and prognosis of cancer.[ This has encouraged researchers to investigate the effect of psychological factors (PFs) on the initiation and prognosis of cancer.[ As a result, several studies have been published on the interactions between cancer and psychological factors such as chronic stress, anxiety, distress, depression, and psycho-social support.[ Although evidence of the positive influence of PFs in cancer survival is modest and findings are inconsistent, strong evidence has been obtained regarding the link between cancer progression and factors such as chronic stress, depression, and social isolation.[ According to Straub and Yan, PFs (stress, anxiety, depression) affect the tumor microenvironment (peripheral immune cells and inflammatory processes) via the hypothalamic–pituitary–adrenal axis, the sympathetic nervous system, and non-adrenal stress hormones, which may alter disease prognosis.[

Many randomized controlled trails have examined the relationship between PFs and the immune system in cancer.[ Most of these trials have focused on the effect of psychological intervention (PI) on immune function. These PIs mainly include cognitive behavior therapy (CBT), stress management (SM), mind–body therapy (MT), and psychological support (PS), while the immune indicators mostly involved are the counts of immune cells, cytokines, and activity of NK cells. Although several meta-analyses have been conducted to collate the evidence regarding the effects of PI on immune response,[ systematic analysis of the effects of different PIs at different stages in cancer treatment on immune function is generally lacking. In this study, we sought to analyze and compare the effect of various PIs administered at different stages of cancer treatment on immune response; we also aimed to evaluate the links between these changes and immune response of cancer patients and possibly their prognosis. We believe that our findings would provide some insights into the psychoneuroimmunology of cancer.

Method

Inclusion and exclusion criteria

The protocol for the meta-analysis was developed in accordance with the PICOS approach. Studies were included in this analysis if they met the following criteria:

randomized controlled trials,

published in Chinese or English,

published before May 2018,

diagnosis of epithelial cancers established according to internationally accepted guidelines,

comparison of PI with usual care, and

outcomes recorded as post-treatment changes in immunological parameters.

Studies were excluded from the analysis if

they were not published in English or Chinese,

patients had any immunological or psychological diseases,

patients had received immune therapy for cancer or drugs for mental illness; and

the study design was other than randomized controlled trail.

The complete details about our study protocol are provided in the About pages at http://www.crd.york.ac.uk/ PROSPERO. The study is a meta-analysis which did not involve any interest of cancer patients, so the ethical review is not necessary.

Search strategy

A systematic computer-based literature search was conducted using relevant databases, including the Cochrane Library, EMBASE, PubMed, Web of Science, Chinese Biomedical Literature Database, Chinese Journal Full-Text Database, VIP Database, and Wanfang Database. We used the following search terms: “cancer” or “tumor” or “tumors” or “tumours” or “carcinoma” or “neoplasm” or “neoplasms” or “oncology” or “oncological”; and “psychological” or “psychology” or “emotion” or “psychotherapy”; and “recovery” or “reduce” or “therapy” or “ treatment” or “therapeutical” or “support” or “counsel”; and “immune” or “immunology”; and “immunological” and “random controlled trials” or “random.”

Study selection and data extraction

After eliminating duplicates using EndNote X7, the title, keywords, abstracts, and contents of all the articles retrieved were independently screened by two reviewers to check if they met the inclusion criteria. If there was any disagreement or doubt about potentially relevant articles, three reviewers jointly decided whether or not the study should be included in this review. Two independent reviewers extracted the data from each study, including authors, year of publication, type and stage of cancer, size of sample, mean patient age, intervention method, type of adjuvant treatment, duration of intervention, and immune outcome.

Data analysis/synthesis

We used Review Manager 5.3 (Cochrane Collaboration, Oxford, United Kingdom) for the meta-analysis. Since the parameters for the measurement of immune status were continuous data, the mean and standard deviation were used to collate the results of the studies. Heterogeneity was tested for all combined results by means of a Q statistic (calculated using a chi-square test), and inconsistency was calculated using an I2 index to determine the impact of heterogeneity. The presence of significant heterogeneity suggests diversity in the various characteristics of the studies, including stage of disease, age, diagnosis, gender, setting, intervention time, and type of assay. When the heterogeneity test was not statistically significant (I2 < 60%, P > .05), a fixed model was used; otherwise, a random effect model or subgroup analysis was used. However, when the heterogeneity of a subgroup analysis was still high (I2 > 60%, P < .05), the random effect model was used.

Literature quality analysis

Two independent reviewers assessed the internal validity of the studies using Cochrane Collaboration's tool (CCT) for assessing risk of bias Any disagreements were resolved by consultation with a third reviewer. The CCT[ is an effective instrument for the evaluation of the internal validity of randomized controlled trials. The quality of a study was classified as strong, moderate, or weak on the basis of the following six domains:

selection bias: random sequence generation and allocation concealment;

performance bias: blinding of participants and personnel;

detection bias: blinding of outcome assessment;

attrition bias: incomplete outcome data;

reporting bias: selective outcome reporting; and

other bias.

If the study was without bias, it was considered to be of high quality; if there was some literature bias, it was deemed to be of moderate quality; and if there was evidence of all types of bias, the study was classified as being of poor quality.

Results

Study selection

After removal of duplicates using EndNote X7, and screened for title and their data abstracted by the inclusion criteria, 29 publications were finally included in this review (Fig. 1).

Figure 1

Flow chart of the study selection process.

Study characteristics, publication bias, and quality of studies

Twenty-nine studies were included in the meta-analysis, including 17 English studies and 12 Chinese studies. In all studies, the cytokine concentrations were reported in picograms per milliliter (pg/mL). The type of intervention varied across the studies: 7 trails used cognitive behavior therapy; 4 utilized stress management; 8 employed mind-body therapy; and the remaining 10 trails adopted psychological supports. The trials also differed in terms of the cancer treatment period during which PI was administered. In four of the studies, patients received PI after completing therapy; in 6, during chemotherapy (CT); in 3, during radiotherapy (RT); in 12, during surgical treatment (ST); and in 4, during adjunctive (multiple) therapy. Among the included studies, 15 provided data on breast cancer. The characteristics of the 29 included studies are summarized in Table 1 .

Table 1

Characteristics of included studies.

Fifteen of these studies were of high quality, while 14 were of moderate quality. All the included studies reported random sequence generation using methods such as random numbers table, coin tossing, and dice throwing, and they provided complete data and results. Nine studies did not provide details regarding allocation concealment, while 14 studies did not provide a clear description about the blinding of the outcome assessment. Data on publication bias and quality of the studies included are detailed in Table 2.

Table 1 (Continued)

Characteristics of included studies.

Meta-analysis results

The effect of different PI approaches on immune cells

Compared with the control group, the SM group did not show any significant differences in CD3+ cell, CD4+ cell, and CD8+ cell counts; CD4+/CD8+ ratio; or NK cell count (P > .05), although significant changes were noted in the CBT group, MT group, and PS group (P < .05). Compared with MT and PS, the CBT group showed the highest magnitude of immune effect, and only the CBT group showed changes in NK cell activity (Table 3, Figs. 2–5).

Table 2

Publication bias and quality of included studies.

Figure 2

Meta-analysis forest map (A) and funnel plot (B) of the effect of cognitive behavior therapy on immune indicators in cancer patients.

Figure 5

Meta-analysis forest map (A) and funnel plot (B) of the effect of psychological support on immune indicators in cancer patients.

Meta-analysis forest map (A) and funnel plot (B) of the effect of stress management on immune indicators in cancer patients.

Meta-analysis forest map (A) and funnel plot (B) of the effect of mind-body therapy on immune indicators in cancer patients.

The influence of PI on immune cells over various cancer treatment periods

Patients who received PI after cancer treatment completion or during CT did not exhibit changes in the counts of any immune indicators, as compared to the control group (P > .05). However, the counts of CD3+ cell, CD4+ cell, and CD8+ cell counts; CD4+/CD8+ ratio, and NK cell count of patients receiving PI during ST, RT, or adjunctive therapy were significantly different compared with the control group(P < .05, Table 4, Figs. 6–10)

Table 3

Effect sizes of PI on immune indicators according to PI types.

Figure 6

Meta-analysis forest map (A) and funnel plot (B) of the effect of PI during surgery period on immune indicators in cancer patients.

Figure 10

Meta-analysis forest map (A) and funnel plot (B) of the effect of PI during completed therapy period on immune indicators in cancer patients.

Meta-analysis forest map (A) and funnel plot (B) of the effect of PI during chemotherapy period on immune indicators in cancer patients.

Meta-analysis forest map (A) and funnel plot (B) of the effect of PI during radiotherapy period on immune indicators in cancer patients.

Meta-analysis forest map (A) and funnel plot (B) of the effect of PI during adjunctive therapy period on immune indicators in cancer patients.

The influence of PI on immune cells in breast cancer patients

Since many of the included studies focused on the effect of PI in breast cancer patients, we conducted a subgroup analysis for breast cancer patients. The CD3+ cell count, CD4+/CD8+ ratio, and NK cell count in breast cancer patients were significant higher in the PI group than in the control group (P > .05), but there were no differences in the CD4+ cell and CD8+ cell count between the two groups (P > .05, Table 5, Fig. 11).

Table 4

Effect sizes of PI on immune indicators according to treatment types.

Figure 11

Meta-analysis forest map (A) and funnel plot (B) of the effect of PI on immune indicators in breast cancer patients.

The effect of PI on immune cytokines

Compared to patients not receiving PI, those who received PI had significantly higher serum levels of IL-2, IL-4, IFN-γ, lgA, and lgG. However, the differences in the serum levels of IL-6, IL-10, TNF-α, and IgM were not significant (P > .05, Fig. 12).

Figure 12

Meta-analysis forest map (A) and funnel plot (B) of the effect of PI on cytokines and immunoglobulins in cancer patients.

Meta-analysis of heterogeneity

Although we performed a subgroup meta-analysis according to the different PI methods employed, different stages of treatment during which PI was administered, and some of the cancer types, there still exist some heterogeneity. The source of heterogeneity may be attributed to sample size, intervention dosage, cancer stages, and patient characteristics.

Discussion

Different immune effect of different PIs

Although there are many factors that affect cancer patient immunity, studies on psychoneuroimmunology (PNI) have proven that immunomodulation through stressors is a reliable and replicable phenomenon.[ The results of our meta-analysis suggest that no significant immune changes were obtained through SM. To our knowledge, SM is an effective stress-reducing PI. However, the degree of cancer patient participation, compliance, and individual stress levels influences its efficacy; moreover, none of the studies that focused on SM took this point into consideration, and SM intervention showed no significant psychological effect as compared to control analogues.[ To the best of our knowledge, the effect of PI on the immune response may be associated with improvements in psychological emotions, hypothalamic–pituitary–adrenal axis (HPA axis), and the sympathetic nervous system. The reason for the nonsignificant immune effect of SM might be the ineffective nature of the PI or low level of emotion distress.[ The other PI-mediated immune responses may likely be attributed to psychological stress-reduction.[ CBT appears to be the best therapeutic strategy for reducing stress and negative emotions.[ Working through stressful experiences can change a person's individual appraisal of subsequent stressors from a sense of threat to a sense of challenge. Perception of a potential stressor as a challenge may lead to changes and support improved immune function.[ Therefore, the magnitude of CBT influence on the immune response is greater than that of the other three PIs.

PI immune influence over cancer treatment progression

Because cancer patients might receive psychopharmacological treatment and anti-cancer therapy may affect immune cells, we conducted a subgroup analysis on various therapies. Our meta-analysis revealed that PI intervention changed the concentration of T and NK cells in cancer patients when administered during ST, RT, and adjunctive therapy (P < .05), but not after completion of the cancer treatment (P > .05). There were no significant differences in the activity levels of the NK cells between the PI group in the chemotherapy and the control groups. We believe that cancer treatment may affect the concentration of immune cells. Lengacher et al showed that compared to T cells, NK cells were more susceptible to suppression during cancer treatment.[ However, studies still indicate that PI can result in changes in the levels of some immune indicators in cancer patients during different treatment periods.[ Wang et al have shown that NK cell activity is associated with the severity of anxiety and depression in cancer patients and that the degree of psychological recovery might affect NK cell activity.[ However, studies on PI during chemotherapy did not indicate any psychological changes after intervention.[ The lack of significant changes in NK cell activity during CT may be due to unclear psychological PI or immunosuppression effects caused by CT.

The Immune response to PI in breast cancer patients

Our meta-analysis consistently showed that PI can change the CD3+ cell count, CD4+/CD8+ratio, and NK cells in breast cancer patients (P < .05), but not the CD4+ cell and CD8+cell counts (P > .05). CD3+ cells could positively promote and enhance the immune response.[ When the concentrations of CD3+ cells and CD4+/CD8+ increase in breast cancer patients, relapse or metastasis may occur, leading to poor prognoses.[ Therefore, PI may be beneficial to the prognosis of breast cancer patients.

Post-PI influence on immune indicator levels and/or activity and ultimate cancer prognosis.

With respect to the immune response trends, we found that there was an increase or decrease in the T-cells counts, but consistent increases in the NK cell count and activity were observed (P < .05). Four of the 10 studies on NK cells confirmed that PI can improve the NK cell content in cancer patients. Likewise, the overall meta-analysis revealed an increase in NK cell count. Three of five studies on NK cell activity indicated that PI may promote the activity of these cells and the overall meta-analysis revealed an increase in NK cell activity. NK cells, which are members of the innate immune cells family,[ are the first line of defence against tumors and infection, assuming the function of immune surveillance cancer cells direct killing.[ NK cell activity can control the growth and spread of pathogens and tumors, both of which play an active immune-monitoring role in controlling the occurrence and metastasis of primary tumors.[ The concentration and activity of NK cells in cancer patients are generally low[; however, increases in their numbers have a positive influence in terms of enhancing immune surveillance and tumor occurrence prevention, and metastasis.[ Therefore, increases in NK cell count and NK cell activity could have a positive influence on the immune function and, ultimately the prognosis of cancer patients.

Three of the five studies on IL-2 showed that PI can increase IL-2 concentration and the overall meta-analysis revealed an increase in IL-2 levels. Two of the four studies on IL-4 confirmed that IL-4 content increased significantly after PI and the overall meta-analysis showed an increase in the IL-2 level. Three of the six studies on IFN-γ proved that PI can increase IFN-γ levels, and the overall meta-analysis revealed an increase in the level of IFN-γ. IL-2 and IFN-γ can significantly induce NK cells to produce and enhance antitumor activity,[ and low concentration of these cells in cervical cancer has been shown to predict severe disease.[ IL-4 has the effect of inhibiting the growth of breast tumors.[ Therefore, the increase in the content of IL-2, IL-4, and IFN-γ may have a positive effect on the immune function and prognosis of cancer patients. Two of the three studies on immunoglobulins confirmed that PI could increase the content of IgA and IgG. The immunoglobulin content reduces in patients with worsening, progressive cancer, and poor prognosis.[ The increase in the concentrations of immunoglobulins may have some beneficial effect in the prognosis of cancer patients.

Further investigations are necessary to determine the mechanism and stability of the immune effect of PI.

Recent studies show that the immune effect of PI may be related to the neuroendocrine changes caused by cognitive changes and improvement in the patient's psychological state.[ However, our subgroup meta-analysis revealed that SM and PI administered after the completion of the cancer therapy or during CT did not bring about any change in the levels of the immune indicators in cancer patients. The stability of the immune effect of PI may also be influenced by intervention-related factors such as PI duration time,[ content of PI,[ and effect of PI[ as well as the cancer stage, the type of adjuvant treatment,[ the severity of psychological stress disorder,[ the degree of PI participation,[ and ability for recovery from immunosuppression.[ There is also some evidence on the interactions between PI and immune indicators, but the psychoneuroimmunology mechanism underpinning the influence of PI on the immune system still remains unclear and further investigations are necessary to elucidate these.

Limitations

This study has some limitations. Most of the papers retrieved by our search were of moderate quality, and most of the enrolled cancer patients in the included studies were female. Furthermore, due to the lack of studies focusing on similar patient groups, subgroup analyses based on the duration of PI or immune function indicators could not be performed in this study. Another point worth mentioning is that the plausible ability of cancer cells evading detection by the immune system makes it difficult to conclusively define the benefits of PI on an individual's immune response.

Conclusion

There is some evidence that supports the benefits of PI on some immune indicators and these immune changes benefit the overall immune function in cancer patients, and possibly their prognosis. However, the definitive influence of PI remains vague and cannot be conclusively defined in terms of immune function and prognosis in cancer patients. Moreover, further research is necessary to examine the individual influence of various PI types against different cancer treatments.

Acknowledgment

We gratefully acknowledge Ping Zhang, Xia Li, and Qiyao Wang for the literature search, screen, data extract and data analysis. Ping Zhang was a major contributor to the writing of the manuscript and Professor Mo checked the manuscript.

Author contributions

Conceptualization: Xia Li, Qiyao Wang.

Data curation: Ping Zhang, Xia Li, Qiyao Wang.

Formal analysis: Ping Zhang, Xia Li, Qiyao Wang.

Investigation: Ping Zhang, Xia Li, Qiyao Wang.

Writing – original draft: Ping Zhang.

Writing – review & editing: Mo Lin.

Table 5

Effect sizes of PI on immune indicators in breast cancer.