Journal of Clinical Epidemiology 67 (2014) 1375e1376
LETTERS TO THE EDITOR Clinically meaningful nocebo effect in acupuncture?
References
The conclusion of Kooga et al. [1] that ‘‘the nocebo effect of acupuncture is clinically meaningful and the rate of patients with any adverse event may be a more appropriate indicator of the nocebo effect’’ deserves some comments. First, opposing placebo and nocebo is a misconception. Indeed, conditioning and expectations are the main factors triggering a response, whether positive or negative. It is the cognitive information itself, which produces the response, for both the level and the direction [2]. If you place a placebo into a person’s drink without telling him, it does not work! Second, the selection of the trials is a limitation as subjectivity is present in most cases. Kooga et al. analyzed 31 and 39 trials, reporting adverse events and dropouts, respectively. However, allocation concealment was present in only 26 and patients blinding in only 15 (Table 1 of Kooga et al. [1]). Could they provide results when selecting trials, which meet these both basic conditions. Finally, although statistically significant, how could the small difference they observed for adverse events be clinically relevant? Even the small analgesic effect of acupuncture seems to lack clinical relevance and cannot be clearly distinguished from bias [3]. Placebos do not have clinically meaningful objective effects: the subjective patient-reported alleviation is not significant, being observed in only one-third of the trials and only under certain conditions. Placebo is Latin for ‘‘I will please’’; the doctor’s duty is not to please but to help. There is no need for placebos to provide reassurance, comfort, and hope. Several factors are pivotal in establishing and maintaining relationships with patients: patience, openness, attentive listening, trust, sharing authority, and commitment. No placebo can replace them [4]. Alain Braillon Northern Hospital 80000 Amiens, France Tel.: þ33 (0) 322 66 88 83; fax: þ33 (0) 322 66 89 55. E-mail address: [email protected]
Funding: None. Conflict of interest: None. 0895-4356/Ó 2014 Elsevier Inc. All rights reserved.
[1] Koog YH, Lee JS, Wi H. Clinically meaningful nocebo effect in acupuncture treatment: a systematic review. J Clin Epidemiol 2014; 67:858e69. [2] Marchand S, Gaumond I. Placebo and nocebo: how to enhance therapies and avoid unintended sabotage to pain treatment. Pain Manag 2013;3:285e94. [3] Madsen MV, Gøtzsche PC, Hrobjartsson A. Acupuncture treatment for pain: systematic review of randomised clinical trials with acupuncture, placebo acupuncture, and no acupuncture groups. BMJ 2009;338: a3115. [4] Braillon A. Placebo is far from benign: it is disease-mongering. Am J Bioeth 2009;9:36e8. http://dx.doi.org/10.1016/j.jclinepi.2014.08.006
Response to letter: Nocebo and placebo effects may not be separate in acupuncture We thank Dr. Alain Braillon for his interest in our article on the nocebo effect in acupuncture [1]. In our article [2], we calculated the magnitude of the nocebo effect in acupuncture using data from clinical trials and found that this effect is clinically meaningful, with a risk difference of 0.049 [95% confidence interval (CI): 0.012, 0.086]. However, Dr. Alain Braillon argues that, although placebo and nocebo are not separate effects, we took a different attitude toward these two effects in our study [1]. We believe that this argument has arisen from his misconception of our article. Theoretically, patients receiving an inert substance should not present any change in symptoms. Nevertheless, in randomized placebocontrolled trials in which allocated treatments are concealed, patients treated with a placebo often show positive health outcomes and sometimes complain of negative outcomes. Clearly, these responses do not occur without patients’ cognition. In this respect, we agree with the opinion of Dr. Alain Braillon that ‘‘it is the cognitive information . which produces the response’’ [1]. Of these responses, we focused on the negative ones and found a significant nocebo effect in acupuncture. Besides considering previous findings of a placebo effect with acupuncture
Funding: None to declare. Conflict of interest: None to declare.
1376
Letters to the Editor / Journal of Clinical Epidemiology 67 (2014) 1375e1376
[3e7], it can be argued that this treatment has double-sided effects, suggesting that both effects are not separate. Dr. Alain Braillon has also raised a concern that trial selection and quality ratings were conducted on a subjective basis [1]. However, it should be noted that this subjectivity is prevalent across reviews. For example, when a study similar to ours in terms of methodology was published by Madsen et al. [3], it was criticized for not including additional eligible trials [8]. Nonetheless, we are confident that the trials included in our study meet the requisite standards, as we have referenced a number of relevant studies [3e7,9]. Similarly, according to the Cochrane Handbook, reviewing authors may ‘‘reach different judgements about risk of bias’’ [10]. Therefore, to minimize bias, we referenced previous relevant studies [3,4,9]. Besides, when we calculated the nocebo effect across nine acupuncture trials with adequate allocation concealment and sufficient patient blinding, our finding was significant at 0.081 (95% CI: 0.020, 0.141). Finally, Dr. Alain Braillon indicated that the standardized mean difference for the placebo effect of acupuncture was 0.42 (95% CI: 0.23, 0.60) [3]. He argued that our finding of 0.049 for the nocebo effect has little clinical meaning because this small difference appears to lack clinical relevance [1]. However, it should be noted that this value is a risk difference. When we calculated a number needed to harm to seek the clinical relevance, it was 20 (95% CI: 12, 83), meaning that 20 patients needed to be treated to observe one additional adverse event. Is not the value of 20 patients clinically meaningful? Although the placebo effect has been a major theme for decades, the nocebo effect has been insufficiently studied. Expanding on our present study [2], we are now planning research on whether the nocebo effect in general is clinically meaningful. With this aim, retrieving all eligible trials is essential; otherwise, there will be serious bias (also known as location bias). However, this process is difficult and time consuming. Indeed, in a previous well-known study on the placebo effect [9], 13 acupuncture-relevant trials were overlooked [5]. Even when we used a different search strategy and identified trials in PubMed, we found 12 additional trials published before 1985. We believe that this research will disclose the general nocebo effect occurring in clinical situations.
Yun Hyung Koog* Jin Su Lee Hyungsun Wi Honam Research Center Medifarm Hospital YangYul-gil 306 Suncheon 540-300 Republic of Korea *Corresponding author. Tel.: þ82-61-729-3300; fax: þ82-61-729-3363. E-mail address: [email protected] (Y.H. Koog)
References [1] Braillon A. Clinically meaningful nocebo effect in acupuncture? J Clin Epidemiol 2014;67:1375. [2] Koog YH, Lee JS, Wi H. Clinically meaningful nocebo effect occurs in acupuncture treatment: a systematic review. J Clin Epidemiol 2014;67:858e69. [3] Madsen MV, Gøtzsche PC, Hrobjartsson A. Acupuncture treatment for pain: systematic review of randomised clinical trials with acupuncture, placebo acupuncture, and no acupuncture groups. BMJ 2009;338:a3115. [4] Linde K, Niemann K, Schneider A, Meissner K. How large are the nonspecific effects of acupuncture? A meta-analysis of randomized controlled trials. BMC Med 2010;8:75. [5] Koog YH, We SR, Min BI. Three-armed trials including placebo and no-treatment groups may be subject to publication bias: systematic review. PLoS One 2011;6:e20679. [6] We SR, Koog YH, Park MS, Min BI. Placebo effect was influenced by publication year in three-armed acupuncture trials. Complement Ther Med 2012;20:83e92. [7] Koog YH, Jung WY. Time course of placebo effect of acupuncture on pain: a systematic review. ISRN Pain 2013;2013:204108. [8] Lao L. Acupuncture treatment for pain: trial selection and threshold concerns. Available at http://www.bmj.com/content/338/bmj.a3115?sort_ byZfield_highwire_a_epubdate_value&sort_orderZDESC&items_ per_pageZ10&pageZ1&panels_ajax_tab_tabZbmj_related_rapid_ responses. Accessed July 23, 2014. [9] Hrobjartsson A, Gøtzsche PC. Placebo interventions for all clinical conditions. Cochrane Database Syst Rev 2010;CD003974. [10] Higgins JPT, Altman DG, Sterne JAC, on behalf of the Cochrane Statistical Methods Group and the Cochrane Bias Methods Group. 8.3.4 Collecting information for assessments of risk of bias. In: Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions version 5.1.0 [updated March 2011]. Available at http://www. cochrane-handbook.org. Accessed July 26, 2014. http://dx.doi.org/10.1016/j.jclinepi.2014.08.005
LETTERS TO THE EDITOR Clinically meaningful nocebo effect in acupuncture?
References
The conclusion of Kooga et al. [1] that ‘‘the nocebo effect of acupuncture is clinically meaningful and the rate of patients with any adverse event may be a more appropriate indicator of the nocebo effect’’ deserves some comments. First, opposing placebo and nocebo is a misconception. Indeed, conditioning and expectations are the main factors triggering a response, whether positive or negative. It is the cognitive information itself, which produces the response, for both the level and the direction [2]. If you place a placebo into a person’s drink without telling him, it does not work! Second, the selection of the trials is a limitation as subjectivity is present in most cases. Kooga et al. analyzed 31 and 39 trials, reporting adverse events and dropouts, respectively. However, allocation concealment was present in only 26 and patients blinding in only 15 (Table 1 of Kooga et al. [1]). Could they provide results when selecting trials, which meet these both basic conditions. Finally, although statistically significant, how could the small difference they observed for adverse events be clinically relevant? Even the small analgesic effect of acupuncture seems to lack clinical relevance and cannot be clearly distinguished from bias [3]. Placebos do not have clinically meaningful objective effects: the subjective patient-reported alleviation is not significant, being observed in only one-third of the trials and only under certain conditions. Placebo is Latin for ‘‘I will please’’; the doctor’s duty is not to please but to help. There is no need for placebos to provide reassurance, comfort, and hope. Several factors are pivotal in establishing and maintaining relationships with patients: patience, openness, attentive listening, trust, sharing authority, and commitment. No placebo can replace them [4]. Alain Braillon Northern Hospital 80000 Amiens, France Tel.: þ33 (0) 322 66 88 83; fax: þ33 (0) 322 66 89 55. E-mail address: [email protected]
Funding: None. Conflict of interest: None. 0895-4356/Ó 2014 Elsevier Inc. All rights reserved.
[1] Koog YH, Lee JS, Wi H. Clinically meaningful nocebo effect in acupuncture treatment: a systematic review. J Clin Epidemiol 2014; 67:858e69. [2] Marchand S, Gaumond I. Placebo and nocebo: how to enhance therapies and avoid unintended sabotage to pain treatment. Pain Manag 2013;3:285e94. [3] Madsen MV, Gøtzsche PC, Hrobjartsson A. Acupuncture treatment for pain: systematic review of randomised clinical trials with acupuncture, placebo acupuncture, and no acupuncture groups. BMJ 2009;338: a3115. [4] Braillon A. Placebo is far from benign: it is disease-mongering. Am J Bioeth 2009;9:36e8. http://dx.doi.org/10.1016/j.jclinepi.2014.08.006
Response to letter: Nocebo and placebo effects may not be separate in acupuncture We thank Dr. Alain Braillon for his interest in our article on the nocebo effect in acupuncture [1]. In our article [2], we calculated the magnitude of the nocebo effect in acupuncture using data from clinical trials and found that this effect is clinically meaningful, with a risk difference of 0.049 [95% confidence interval (CI): 0.012, 0.086]. However, Dr. Alain Braillon argues that, although placebo and nocebo are not separate effects, we took a different attitude toward these two effects in our study [1]. We believe that this argument has arisen from his misconception of our article. Theoretically, patients receiving an inert substance should not present any change in symptoms. Nevertheless, in randomized placebocontrolled trials in which allocated treatments are concealed, patients treated with a placebo often show positive health outcomes and sometimes complain of negative outcomes. Clearly, these responses do not occur without patients’ cognition. In this respect, we agree with the opinion of Dr. Alain Braillon that ‘‘it is the cognitive information . which produces the response’’ [1]. Of these responses, we focused on the negative ones and found a significant nocebo effect in acupuncture. Besides considering previous findings of a placebo effect with acupuncture
Funding: None to declare. Conflict of interest: None to declare.
1376
Letters to the Editor / Journal of Clinical Epidemiology 67 (2014) 1375e1376
[3e7], it can be argued that this treatment has double-sided effects, suggesting that both effects are not separate. Dr. Alain Braillon has also raised a concern that trial selection and quality ratings were conducted on a subjective basis [1]. However, it should be noted that this subjectivity is prevalent across reviews. For example, when a study similar to ours in terms of methodology was published by Madsen et al. [3], it was criticized for not including additional eligible trials [8]. Nonetheless, we are confident that the trials included in our study meet the requisite standards, as we have referenced a number of relevant studies [3e7,9]. Similarly, according to the Cochrane Handbook, reviewing authors may ‘‘reach different judgements about risk of bias’’ [10]. Therefore, to minimize bias, we referenced previous relevant studies [3,4,9]. Besides, when we calculated the nocebo effect across nine acupuncture trials with adequate allocation concealment and sufficient patient blinding, our finding was significant at 0.081 (95% CI: 0.020, 0.141). Finally, Dr. Alain Braillon indicated that the standardized mean difference for the placebo effect of acupuncture was 0.42 (95% CI: 0.23, 0.60) [3]. He argued that our finding of 0.049 for the nocebo effect has little clinical meaning because this small difference appears to lack clinical relevance [1]. However, it should be noted that this value is a risk difference. When we calculated a number needed to harm to seek the clinical relevance, it was 20 (95% CI: 12, 83), meaning that 20 patients needed to be treated to observe one additional adverse event. Is not the value of 20 patients clinically meaningful? Although the placebo effect has been a major theme for decades, the nocebo effect has been insufficiently studied. Expanding on our present study [2], we are now planning research on whether the nocebo effect in general is clinically meaningful. With this aim, retrieving all eligible trials is essential; otherwise, there will be serious bias (also known as location bias). However, this process is difficult and time consuming. Indeed, in a previous well-known study on the placebo effect [9], 13 acupuncture-relevant trials were overlooked [5]. Even when we used a different search strategy and identified trials in PubMed, we found 12 additional trials published before 1985. We believe that this research will disclose the general nocebo effect occurring in clinical situations.
Yun Hyung Koog* Jin Su Lee Hyungsun Wi Honam Research Center Medifarm Hospital YangYul-gil 306 Suncheon 540-300 Republic of Korea *Corresponding author. Tel.: þ82-61-729-3300; fax: þ82-61-729-3363. E-mail address: [email protected] (Y.H. Koog)
References [1] Braillon A. Clinically meaningful nocebo effect in acupuncture? J Clin Epidemiol 2014;67:1375. [2] Koog YH, Lee JS, Wi H. Clinically meaningful nocebo effect occurs in acupuncture treatment: a systematic review. J Clin Epidemiol 2014;67:858e69. [3] Madsen MV, Gøtzsche PC, Hrobjartsson A. Acupuncture treatment for pain: systematic review of randomised clinical trials with acupuncture, placebo acupuncture, and no acupuncture groups. BMJ 2009;338:a3115. [4] Linde K, Niemann K, Schneider A, Meissner K. How large are the nonspecific effects of acupuncture? A meta-analysis of randomized controlled trials. BMC Med 2010;8:75. [5] Koog YH, We SR, Min BI. Three-armed trials including placebo and no-treatment groups may be subject to publication bias: systematic review. PLoS One 2011;6:e20679. [6] We SR, Koog YH, Park MS, Min BI. Placebo effect was influenced by publication year in three-armed acupuncture trials. Complement Ther Med 2012;20:83e92. [7] Koog YH, Jung WY. Time course of placebo effect of acupuncture on pain: a systematic review. ISRN Pain 2013;2013:204108. [8] Lao L. Acupuncture treatment for pain: trial selection and threshold concerns. Available at http://www.bmj.com/content/338/bmj.a3115?sort_ byZfield_highwire_a_epubdate_value&sort_orderZDESC&items_ per_pageZ10&pageZ1&panels_ajax_tab_tabZbmj_related_rapid_ responses. Accessed July 23, 2014. [9] Hrobjartsson A, Gøtzsche PC. Placebo interventions for all clinical conditions. Cochrane Database Syst Rev 2010;CD003974. [10] Higgins JPT, Altman DG, Sterne JAC, on behalf of the Cochrane Statistical Methods Group and the Cochrane Bias Methods Group. 8.3.4 Collecting information for assessments of risk of bias. In: Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions version 5.1.0 [updated March 2011]. Available at http://www. cochrane-handbook.org. Accessed July 26, 2014. http://dx.doi.org/10.1016/j.jclinepi.2014.08.005
