Jump to comment:

• Response to "Depression and anxiety in ovarian cancer: a systematic review and meta-analysis of prevalence rates"
  Associate Professor Vanessa L Beesley
  Published on: 17 December 2020

• Published on: 17 December 2020

  Response to "Depression and anxiety in ovarian cancer: a systematic review and meta-analysis of prevalence rates"

  • Associate Professor Vanessa L Beesley, Associate Professor QIMR Berghofer Medical Research Institute

  We read with interest the systematic review and meta-analysis by Watts et al. published in BMJ Open [1] which reported on the prevalence of depression and anxiety in women with ovarian cancer at three time-points: pre-treatment, on-treatment and post-treatment. We agree this is an important topic but, after reviewing the article, would like to raise some concerns. We have reproduced the key components of Table 1 from that paper with additional comments to note some potential inaccuracies. Our main concerns are as follows:
  First, it appears that the same women have been counted twice in some analyses. The most concerning instance of this relates to the Australian Ovarian Cancer Study, which is by far the largest study with 794 cases with data on depression and anxiety. Women in this study have been double counted (Price 2009 [2] and Price 2010 [3]) in the pre-treatment analysis, then also counted (Price 2009) in the on-treatment analysis, when in practice 79% of women in this study were post-treatment so these data (one or other of the reports) should have been included in the post-treatment analyses only (but they are not). Watts et al. also include two papers (Liavaag 2007 [4] and Liavaag 2009 [5]) from a single study at the Norwegian Radium Hospital and count these women in two separate analyses. Furthermore, they include four papers from a group with study sites in Iowa, Miami and Texas (Lutgendorf 2008 [6], Lutgendorf 2008 [7], Lutgendorf 2009 [8] and Clevenger 2...

  Show More

  We read with interest the systematic review and meta-analysis by Watts et al. published in BMJ Open [1] which reported on the prevalence of depression and anxiety in women with ovarian cancer at three time-points: pre-treatment, on-treatment and post-treatment. We agree this is an important topic but, after reviewing the article, would like to raise some concerns. We have reproduced the key components of Table 1 from that paper with additional comments to note some potential inaccuracies. Our main concerns are as follows:
  First, it appears that the same women have been counted twice in some analyses. The most concerning instance of this relates to the Australian Ovarian Cancer Study, which is by far the largest study with 794 cases with data on depression and anxiety. Women in this study have been double counted (Price 2009 [2] and Price 2010 [3]) in the pre-treatment analysis, then also counted (Price 2009) in the on-treatment analysis, when in practice 79% of women in this study were post-treatment so these data (one or other of the reports) should have been included in the post-treatment analyses only (but they are not). Watts et al. also include two papers (Liavaag 2007 [4] and Liavaag 2009 [5]) from a single study at the Norwegian Radium Hospital and count these women in two separate analyses. Furthermore, they include four papers from a group with study sites in Iowa, Miami and Texas (Lutgendorf 2008 [6], Lutgendorf 2008 [7], Lutgendorf 2009 [8] and Clevenger 2013 [9]) and there is likely substantial overlap in the participants included in these papers (Professor Lutgendorf, personal communication).
  Second, there appears to be some misclassification of the studies contributing to each analysis of treatment phase. In addition to the misclassification of Price et al. (mentioned above), it appears data from Hipkins 2004 [10] were included in the pre-treatment meta-analysis, when women in this study completed measures at completion of chemotherapy (T1) and 3 months after chemotherapy (T2). Data from Meraner 2004 [11] were also included in the pre-treatment meta-analysis, when baseline data in this study were collected after surgery at first chemotherapy cycle which could be considered on-treatment, particularly for psychological symptoms. Data from Liavaag 2007 [4] were included in the on-treatment analysis, when women in this study were at least 18 months post-diagnosis (n=130 non-relapsed and n=59 relapsed) and thus most would have been post-treatment. Data from Clevenger 2013 [9] were included in the on-treatment analysis, when women in this study completed measures before surgery (T1) and at 6 months (T2) and 1 year after diagnosis (T3) and thus are unlikely to have been on treatment at any time-point. The post-treatment results for depression that were included from Goncalves 2010 [12] are from women who are 3 months post-diagnosis and thus likely to be on-treatment. In contract, data from Goncalves 2008 [13] collected 3 months after cessation of treatment have not been included in the post-treatment analysis despite being available. Furthermore, the results for anxiety and depression on-treatment from this paper appear to have been reversed. Data from Stafford 2010 (corrected year 2011) [14] were included in the on-treatment analysis when only 14% of the 71 women with ovarian cancer were currently receiving treatment at the time of the survey (the remained were post-treatment). Additionally, data for depression from Stafford 2010/11 [14] were omitted from the post-treatment analysis.
  Third, there is variation in the thresholds used to classify depression and anxiety (shown in Table 1 linked below).
  Notably, data from Price (2009 [2] and 2010 [3]), Goncalves 2010 [12], Bisseling 2009 [15], and Hodgkinson 2007 [16] use the clinical threshold (HADS ≥11), while Liavaag (2007 [4] and 2009 [5]), Goncalves 2008 [13], and Hipkins 2004 [10] have used the subclinical + clinical threshold (HADS ≥8) to define cases and Urbaniec (2011) [17], Sukegawa (2008) [18] and Schulman-Green (2008) [19] use STAI cut-offs of 40, 42 and 48 respectively. Thus the range in the prevalence results between studies reflects, in part, the different thresholds used. Moreover, Sukegawa (2008) [18], Schulman-Green (2008) [19], Meraner (2012) [11], Holzner (2003) [20] and Stafford (2011) [14] only presented means and standard deviations for anxiety and depression. Watts et al. does not specify how they calculated prevalence for these studies nor what thresholds they used in their calculations.
  We also noted a number of other minor inaccuracies in Table 1. For example, there are some inaccuracies in reported study size which may have affected the calculated confidence intervals for the prevalence estimates. Furthermore, Urbaniec 2011 [17] only reported anxiety and depression scores from a mix of gynaecological cancer types rather than ovarian cancer specifically with no indication if there were differences by cancer type. On this basis, we suggest this study should not have been included.
  These concerns raise questions about the reliability of the prevalence estimates presented. We decided to recalculate the pooled results. To do this, we (a) removed duplicate results (and in Professor Lutgendorf papers where there was likely more than 50% overlap we selected the most recent paper with the most extensive recruitment), (b) reallocated studies to the correct treatment phase as discussed above, (c) used the HADS ≥8 threshold, which is recommended in cancer populations [21] and STAI ≥40 which is most commonly used, [22] and (d) used Z scores to estimate the proportion over the threshold value of interest (HADS=8; STAI=40; CES-D=16) when only means scores were available or a different threshold was used and the mean score was available. We have reproduced Figures 2, Figure 3 and Figure 4 (linked below) from the original paper and estimate that the true pooled estimates for depression prevalence pre-treatment was 47% (CI 42-53%), on-treatment was 29% (CI 25-33%) and post-treatment was 15% (CI 13-17%). For anxiety, the corresponding figures were pre-treatment 54% (CI 45-63%), on-treatment 43% (CI 36-50%) and post-treatment 26% (CI 24-29%). Thus in particular Watts et al’s original estimates for depression pre-treatment (25%; CI 23-28%) and for anxiety pre-treatment (19%; CI 17-21%) and on-treatment (26%; CI 22-31%) appear to be gross underestimates. We suggest our updated pooled estimates are more appropriate reference figures moving forwards.

  Yours Sincerely,
  Associate Professor Vanessa L Beesley, Population Health Department, QIMR Berghofer Medical Research Institute
  Professor Penelope M Webb, Population Health Department, QIMR Berghofer Medical Research Institute

  Table 1:
  https://blogs.bmj.com/bmjopen/files/2020/12/BMJ-Open-Paper-Response-Tabl...

  Figure 2:
  https://blogs.bmj.com/bmjopen/files/2020/12/BMJ-Open-Paper-Response-Figu...

  Figure 3:
  https://blogs.bmj.com/bmjopen/files/2020/12/BMJ-Open-Paper-Response-Figu...

  Figure 4:
  https://blogs.bmj.com/bmjopen/files/2020/12/BMJ-Open-Paper-Response-Figu...

  References
  1. Watts, S., et al., Depression and anxiety in ovarian cancer: a systematic review and meta-analysis of prevalence rates. BMJ Open, 2015. 5(11): p. e007618-e007618.
  2. Price, M.A., et al., Prevalence and predictors of insomnia in women with invasive ovarian cancer: anxiety a major factor. European Journal Of Cancer (Oxford, England: 1990), 2009. 45(18): p. 3262-3270.
  3. Price, M.A., et al., Prevalence and predictors of anxiety and depression in women with invasive ovarian cancer and their caregivers. The Medical Journal Of Australia, 2010. 193(5 Suppl): p. S52-S57.
  4. Liavaag, A.H., et al., Controlled study of fatigue, quality of life, and somatic and mental morbidity in epithelial ovarian cancer survivors: how lucky are the lucky ones? Journal Of Clinical Oncology: Official Journal Of The American Society Of Clinical Oncology, 2007. 25(15): p. 2049-2056.
  5. Liavaag, A.H., et al., Morbidity associated with "self-rated health" in epithelial ovarian cancer survivors. BMC Cancer, 2009. 9: p. 2-2.
  6. Lutgendorf, S.K., et al., Interleukin-6, cortisol, and depressive symptoms in ovarian cancer patients. Journal Of Clinical Oncology: Official Journal Of The American Society Of Clinical Oncology, 2008. 26(29): p. 4820-4827.
  7. Lutgendorf, S.K., et al., Biobehavioral influences on matrix metalloproteinase expression in ovarian carcinoma. Clinical Cancer Research: An Official Journal Of The American Association For Cancer Research, 2008. 14(21): p. 6839-6846.
  8. Lutgendorf, S.K., et al., Depression, social support, and beta-adrenergic transcription control in human ovarian cancer. Brain, Behavior, And Immunity, 2009. 23(2): p. 176-183.
  9. Clevenger, L., et al., Sleep disturbance, distress, and quality of life in ovarian cancer patients during the first year after diagnosis. Cancer, 2013. 119(17): p. 3234-3241.
  10. Hipkins, J., et al., Social support, anxiety and depression after chemotherapy for ovarian cancer: a prospective study. British Journal Of Health Psychology, 2004. 9(Pt 4): p. 569-581.
  11. Meraner, V., et al., Monitoring physical and psychosocial symptom trajectories in ovarian cancer patients receiving chemotherapy. BMC Cancer, 2012. 12: p. 77-77.
  12. Gonçalves, V., G. Jayson, and N. Tarrier, A longitudinal investigation of psychological disorders in patients prior and subsequent to a diagnosis of ovarian cancer. Journal Of Clinical Psychology In Medical Settings, 2010. 17(2): p. 167-173.
  13. Gonçalves, V., G. Jayson, and N. Tarrier, A longitudinal investigation of psychological morbidity in patients with ovarian cancer. British Journal Of Cancer, 2008. 99(11): p. 1794-1801.
  14. Stafford, L. and F. Judd, Long-term quality of life in Australian women previously diagnosed with gynaecologic cancer. Supportive Care In Cancer: Official Journal Of The Multinational Association Of Supportive Care In Cancer, 2011. 19(12): p. 2047-2056.
  15. Bisseling, K.C.H.M., et al., Depression, anxiety and body image after treatment for invasive stage one epithelial ovarian cancer. The Australian & New Zealand Journal Of Obstetrics & Gynaecology, 2009. 49(6): p. 660-666.
  16. Hodgkinson, K., et al., Long-term survival from gynecologic cancer: Psychosocial outcomes, supportive care needs and positive outcomes. Gynecologic Oncology, 2007. 104(2): p. 381-9.
  17. Urbaniec, O.A., et al., Gynecological cancer survivors: assessment of psychological distress and unmet supportive care needs. Journal Of Psychosocial Oncology, 2011. 29(5): p. 534-551.
  18. Sukegawa, A., et al., Anxiety and prevalence of psychiatric disorders among patients awaiting surgery for suspected ovarian cancer. The Journal Of Obstetrics And Gynaecology Research, 2008. 34(4): p. 543-551.
  19. Schulman-Green, D., et al., Quality of life among women after surgery for ovarian cancer. Palliative & Supportive Care, 2008. 6(3): p. 239-247.
  20. Holzner, B., et al., Fatigue in ovarian carcinoma patients: a neglected issue? Cancer, 2003. 97(6): p. 1564-1572.
  21. Vodermaier, A. and R.D. Millman, Accuracy of the Hospital Anxiety and Depression Scale as a screening tool in cancer patients: a systematic review and meta-analysis. Supportive Care In Cancer: Official Journal Of The Multinational Association Of Supportive Care In Cancer, 2011. 19(12): p. 1899-1908.
  22. Julian, L.J., Measures of anxiety: State-Trait Anxiety Inventory (STAI), Beck Anxiety Inventory (BAI), and Hospital Anxiety and Depression Scale-Anxiety (HADS-A). Arthritis care & research, 2011. 63 Suppl 11: p. S467-S472.
  23. Wenzel, L.B., et al., Resilience, reflection, and residual stress in ovarian cancer survivorship: a gynecologic oncology group study. Psychooncology, 2002. 11(2): p. 142-53.
  24. Norton, T.R., et al., Prevalence and predictors of psychological distress among women with ovarian cancer. Journal Of Clinical Oncology: Official Journal Of The American Society Of Clinical Oncology, 2004. 22(5): p. 919-926.
  25. Parker, P.A., et al., The associations between knowledge, CA125 preoccupation, and distress in women with epithelial ovarian cancer. Gynecologic Oncology, 2006. 100(3): p. 495-500.
  26. Costanzo, E.S., et al., Psychosocial factors and interleukin-6 among women with advanced ovarian cancer. Cancer, 2005. 104(2): p. 305-313.
  27. Slovacek, L., et al., Screening for depression in survivors of metastatic ovarian cancer in a programme of palliative cancer care. Bratislavske Lekarske Listy, 2009. 110(10): p. 655-659.

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top