Reflexology in oncological treatment – a systematic review

Abstract

Background

As cancer and its therapy comes with a wide range of negative effects, people look for options to mitigate these effects. Reflexology is among the options of complementary medicine.

Method

In March 2022 a systematic search was conducted searching five electronic databases (Embase, Cochrane, PsychInfo, CINAHL and Medline) to find studies concerning the use, effectiveness and potential harm of reflexology on cancer patients.

Results

From all 821 search results, 29 publications concerning 26 studies with 2465 patients were included in this systematic review. The patients treated with reflexology were mainly diagnosed with breast, lung, gastrointestinal and hematological cancer. Outcomes were mainly pain, quality of life, anxiety, depression, fatigue. The studies had moderate to low quality and reported heterogeneous results: Some studies reported significant improvements in above mentioned outcomes while other studies did not find any changes concerning these endpoints.

Conclusion

Due to the very heterogeneous results and methodical limitations of the included studies, a clear statement regarding the effectiveness of reflexology on cancer patients is not possible. The current evidence indicates that reflexology is superior to passive control groups for pain, quality of life and fatigue, however, more studies with comparable active control groups are needed.

Introduction

Cancer is a very prevalent disease with more than 18 million diagnosed cases worldwide in 2020 [1]. Due to the burden of the disease and adverse effects of cancer treatment, people look for options that might help mitigate these negative effects, with reflexology being a particularly popular option among complementary medicine. Reflexology involves applying manual pressure to specific parts of the body (often feet, sometimes hands) that are thought to correspond with specific internal organs. The stimulation of the body zones is intended to promote the self-healing powers of the organs that are associated with the respective zones. Originally developed as so called “Zone therapy” by William Fitzgerald, Eunice Ingham refined these techniques [2] and her method of reflexology is still used today. These reflex zones are also not to be confused with so called Head zones, named after neurologist Henry Head. He found that visceral diseases can result in hyperalgesia or allodynia of respective cutaneous areas [3].

This review aims at assessing clinical studies on the influence of reflexology as complementary medicine on cancer related symptoms and side effects of cancer therapy. It is not clear yet, whether differences in application might yield different results for a variety of outcomes in the context of cancer patients, which also applies to acute and long-term effects. Additionally, comparisons with other interventions that aim at improving the patients’ condition may help shed more light on the efficacy of reflexology. This exploration may help guide how healthcare practitioners can support cancer patients’ symptoms better and if reflexology can be an adequate tool in doing so.

Method

Criteria for including and excluding studies in the review

Inclusion and exclusion criteria are listed in Table 1 based on a PICO- model. Generally, all original studies with a randomized controlled design were included if they reported patient-relevant outcomes after treatment of adult cancer patients with any intervention containing reflexology. Because of the wide range of application fields, all cancer entities were included. Criteria for rejecting studies were primary prevention, grey literature, other publication type than primary investigation/report (e.g. comments, letters, abstracts) and study population with precancerous conditions. Additionally, studies were excluded if they reported no patient centered outcomes. Language restrictions were made to English and German. In order to shed more light on the effectiveness of reflexology compared to other non-specific interventions a distinction was made between active and passive control groups.

Table 1 PICO criteria

Search and study selection

While searching for studies and selecting them, we followed the approach described in a systematic review by Römer et al. {Römer, 2021 #496}. A systematic research was conducted using five databases (Medline (Ovid), CINAHL (EBSCO), EMBASE (Ovid), Cochrane CENTRAL and PsycINFO (EBSCO)) in March 2022. For each of these databases a complex search strategy was developed consisting of a combination of MeshTerms, keywords and text words in different spellings connected to cancer and reflexology (Table 2). The search string was restricted by filters of study or publication type. After importing the search results into EndNote 20, all duplicates were removed and a title- abstract- screening was carried out by three independent reviewers (MK, JD, SK). In case of disagreement consensus was made by discussion or a fourth reviewer ¹was consulted (JH). Furthermore, systematic reviews, which cover studies with a randomized controlled design were screened for relevant studies. When title and abstract did not have sufficient information for screening purposes, a full-text copy was retrieved as well. After that, all full texts were retrieved and screened again independently by both reviewers. Additionally, bibliography lists of all retrieved articles were searched for relevant studies.

Table 2 Search string reflexology - March 2022

Excluded studies

Excluded were 8 RCTs due to outcomes not being patient-relevant, patients not being cancer patients and multiple interventions. As the effects of the single parts of these interventions are not known and were not analyzed separately, it is not possible to estimate whether the reported effects are caused by the reflexology or by another treatment. A list of excluded studies can be seen in Appendix 1.

Assessment of risk of bias and methodological quality

All characteristics were assessed by two independent reviewers (MK, JD). In case of disagreement a third reviewer was consulted (JH) and consensus was made by discussion.

The risk of bias in the included studies was analyzed with the Cochrane revised Risk of Bias Tool 2.0 [4].

Additional criteria concerning methodology were size of population, application of power analysis, adequacy of statistical tests (e.g. control of premises or multiple testing) and selective outcome reporting (report of all assessed outcomes with specification of statistical data as the p-value).

Data extraction

Data extraction was performed by one reviewer (MK) and controlled by two independent reviewers (JD, JH). As a template for data extraction, the evidence tables from the National Guideline on Complementary and Alternative Medicine in Oncological Patients of the German Guideline Program in Oncology were used.

Results

The systematic search revealed 821 results. No studies were added by hand search. At first, duplicates were removed leaving 479 studies. After screening title and abstract, 133 studies remained to complete review (see Consort diagram, Fig. 1). Finally, 29 publications were analyzed in this review, including 29 RCT. According to this, the 29 publications reported data from 26 relevant studies. Detailed characterization of the included studies may be seen in Table 3.

Fig. 1

PRISMA flow diagram

Table 3 Short Evidence table

Characteristics of included studies

Concerning all relevant studies, 2465 patients were included and 2262 of them were analyzed, due to 405 drop outs. The age of patients ranged from 18 to 98 years. 70.8% of the participants were female. Endpoints these studies investigated include pain, anxiety and depression, fatigue, QoL/symptom severity and distress, physical and social functioning/interference with daily life, nausea and vomiting, sleep, mood, relaxation, narcotic analgesia consumption, self-esteem, psychiatric morbidity, perceived social support and quality of relationship between caregiver and patient. While physical and social functioning/interference with daily life could also be counted towards QoL, for the sake of clarity we decided to report them separately.

Risk of bias in included studies

The methodical quality was assessed with the Cochrane revised Risk of Bias Tool 2.0 [4]. The results are presented in Table 4. Three of the included studies show moderate risk of bias and 26 show high risk of bias.

Table 4 Risk of Bias Assessment

Efficacy of reflexology

Pain

Description of studies

Fifteen RCTs dealt with the effects of reflexology on pain. In eight of these [5,6,7,8,9,10,11,12,13], the intervention was carried out by a certified reflexologist, in four [13,14,15,16], the intervention was carried out by caregivers who were taught how to apply the intervention and in three [17,18,19], the intervention was carried out by the researcher but further information on his qualifications regarding reflexology is missing.

Seven of the studies used an active control group [7, 9, 13,14,15, 17, 18], while the remaining seven used a passive one [5, 6, 8, 10,11,12, 16, 19].

Samancioglu Baglama et al. [15] included 64 patients with mainly hematologic disorders who received either a 60 min reflexology or reading session for 15 days. On the last day of intervention, the reflexology group showed a significantly better result on the VAS than the reading group (4.70 ± 1.55 vs. 6.36 ± 0.99; p = 0.000). In a study by Rambod et al. [5], the intervention was applied over five days and showed significant differences between intervention (2.72 ± 2.30) and control (4.33 ± 3.54;) at the end of the study (n = 72; p = 0.01). Dikmen et al. [18] analyzed 80 patients and already found significant differences for pain severity between groups at baseline (p = 0.001). Significant differences were also found at 3^rd, 8^th and at 12^th week (follow-up) for pain severity and effect on daily life (p’s < 0.017) with the lowest scores found in the reflexology plus relaxation group. Jahani et al. [19] included 84 patients and collected data three days before and three days after a three day intervention, showing a significant group difference, with less pain in the intervention group already one day before the intervention (4.12 ± 2.18 vs. 6.57 ± 2.08; p = 0.001), as well as at day one (no p-value), two (p = 0.001) and three (p = 0.001) after. In a study by Stephenson et al. [14] data was collected from 90 patients before and after a reflexology session measuring pain with the brief pain inventory and the Short Form-McGill Pain Questionnaire (SF-MPQ). A significant difference between groups was found (p = 0.001), showing a bigger mean decrease in score in the intervention (1.1 points) compared to the control group (0.1 points; = 0.12). A subgroup analysis only analyzing the 32 patients with a score > 5 also showed a significant decrease in the intervention group with the decrease in score being even bigger (2.7 points) in the intervention group while the control group only decreased by 0.5 points (p = 0.007, = 0.23). Stephenson et al. [11] published another study, which included 36 patients and the intervention group received two sessions of reflexology 24 h apart. The authors found significantly less pain directly after the intervention in the reflexology group compared to the control group, which received standard care (p < 0.01). However, no such differences were found at three and 24 h after the intervention. Tsay et al. [8] investigated pain in 62 subjects with a VAS and the SF-MPQ applying reflexology on day two to four post-surgery for digestive cancer. Using the VAS, the authors found significantly lower values in the intervention group (β_G = -21.22 (4.93, p < 0.001) on average over all measurement points. Change of pain over time was also significantly different (p = 0.0107) with pain by trend staying the same in the intervention group while it was getting worse in the control group (β_I = -2.41 (1.38)), which also underwent surgery but received only standard care. For the SF-MPQ data were only collected at baseline and follow up at day five and six post-surgery and did not show any significant differences between groups but a decrease in pain in both arms, which over time was significantly stronger in the intervention group (β_I = -3.17 (1.41); p = 0.02). In a study with 40 patients by Anderson et al. [10] patients received one single session, showing a significant improvement on VAS scores from pre [mean = 4, 95% CI = 2.9, 5] to post session [mean = 1.6, 95% CI = 0.9, 2.2] for the intervention group (p < 0.0001) but not for the control group (mean = 3.7 pre and post session) which filled out surveys during a session. However, they did not directly compare the groups for outcomes but compared them regarding time since last pain medication showing no significant differences. Sikorskii et al. [16] in a secondary analysis of a study by Wyatt et al. [20] compared reflexology to a control group which only received calls for symptom assessment. They compared the Patient Reported Outcomes Measurement System (PROMIS) and Legacy measures (a group of questionnaires) for various outcomes at baseline and one week after the intervention. Significant differences between intervention and control were found one week after the intervention for both PROMIS-29 pain severity profile v1.0 (p = 0.04, ES = 0.31) and M.D. Anderson symptom inventory (MDASI) pain severity (p < 0.01, ES = 0.46) with better results in the intervention group. Stephenson et al. [12] investigated pain in 24 patients with breast and lung cancer using a crossover trial. One group received one reflexology session (30 min) and three days of no intervention with a 30 min control session on the last day and the opposite way for the other group. Measurements were taken before and after the first and the last session using three scales. The SF-MPQ showed significantly better results (mean difference = -0.41; p < 0.05) after reflexology compared to after the control session, while the SF-MPQ:PPI (present pain intensity) Scale and the VAS, both measuring pain intensity, did not find such differences. For all three scales only patients with breast cancer were included. Uysal et al. [13] who included 65 patients for five weeks (two interventions weekly) investigated adverse effects and found significantly less pain with grade 2 + in the reflexology group comparing it with control in week four (p = 0.002) and five (p < 0.001).

Four studies did not report any significant differences after six weeks with the Bayly Method [17] or subscales of the brief pain inventory [6], of which Wyatt et al. [9] used the pain intensity subscale in their cross-over trial (reflexology and Swedish massage, four weeks, washout one week). Hodgson et al. [7] also did not find any significant differences between the two study groups for any time point using the checklist of nonverbal pain indicators (CNPI).

Four of the eight studies in which a certified reflexologist applied the intervention showed significant results in favor of the intervention [5, 8, 11, 12], all three in which a caregiver applied the intervention [14,15,16] and two of the three with missing information [18, 19].

Methodical assessment of studies:

In the study by Dikmen et al. [18], the authors only reported p-values and presented results graphically without providing further information, making an interpretation in terms of clinical significance very difficult. Additionally, the enrollment and allocation process are difficult to understand with a huge dropout and no sufficient baseline information exist. Full blinding of the researcher for statistical analysis as stated in the study is impossible as the researcher conducted the allocation. Furthermore, in the study it wasn’t accounted for the same session duration of all interventions. Stephenson et al. [11] and Sikorskii et al. [16] did not provide information on the homogeneity of the groups or lack thereof [17]. Two other studies are either lacking information on dropouts [19] or had a huge dropout [6]. In the study by Tsay et al. [8], there might be an interference of analgesics with the intervention and one study by Stephenson et al. [14] shows risk of a reporting bias as pain was measured with two tools while reporting only one of them without clarifying which one. Stephenson et al. [12] formed mean values means of both groups and not within group, so patients were not their own control anymore in this crossover trial. This incorrect analysis doesn’t allow for interpretation of the results. This applies to two other studies, as well, as the statistical analysis is incomprehensible [7] or only intragroup comparisons were made [10]. The study by Wyatt et al. [9] also shows a risk for sampling bias and reduced reporting as no results were reported comparing the two active groups except for dyspnoe, demonstrating a significant result.

Anxiety and depression

Description of studies

In eight of the studies dealing with anxiety and depression the intervention was delivered by certified reflexologists [8, 9, 12, 13, 21,22,23,24] while in five it was caregivers delivering it [14,15,16, 25, 26]. In three more studies the researchers applied reflexology but no information are given regarding their qualifications [17, 19, 27].

Eight of the studies used an active control group [9, 13,14,15, 17, 22, 23, 25], whereas the other eight used a passive one [8, 12, 16, 19, 21, 24, 26, 27].

Eight RCTs found a significant effect of reflexology on anxiety and depression in cancer patients [12,13,14,15,16,17, 19, 27]. Mantoudi et al. [17] reported a significant difference in change between the reflexology and relaxation group when comparing baseline values with 4^th (p = 0.006, η² = 0.094) and 6^th week (p = 0.001, η² = 0.138) for depression. For anxiety, however, no significant difference in change was found. Göral Türkcü et al. [27] applied reflexology to 62 patients with gynecological cancers over two weeks and found an advantage for the reflexology group two weeks after the end of the intervention for anxiety (p < 0.001) and depression (p < 0.001). Samancioglu Baglama et al. [15] and Stephenson et al. [14] both used a VAS to explore the effects of reflexology on anxiety. Both found significant differences, in favor of the reflexology group at day 15 (p = 0.036) and directly after a one time intervention (p = 0.001, ε^2 = 0.13), respectively. The latter also did a subgroup analysis for patients with anxiety > 5 revealing a significant difference (p = 0.006; = 0.15). In another study by Stephenson et al. [12] significantly better results were observed for anxiety after a reflexology compared to a control session (mean difference = -21.83; p < 0.000). This time, both, breast and lung cancer patients were analyzed. Using the Spielberger State-Trait Anxiety Inventory, Jahani et al. [19] found a significant advantage of the reflexology group on day three after treatment (n = 84; p = 0.04;), while Rezaei et al. [24] did not find a significant difference (n = 74). Still, there are differences that need to be considered. Rezaei et al. [24] merely did a before and after comparison taking place on the same day whereas Jahani et al. [19] had a three day intervention period and collected data only on day three after the intervention period. Furthermore, they had a passive control group whereas Rezaei et al. [24] had a researcher stand at bedside of the control group and no further information are given about their contact. Sikorskii et al. [16] found significant differences in favor of reflexology compared to a control group for depression using the Center for Epidemiologic Studies Depression Scale (short: CES-D; ES = 0.32, p = 0.03), MDASI distress severity (ES = 0.31 and p = 0.04) and SF-36 mental health (ES = 0.51, p < 0.01). Using PROMIS-29 depression profile v1.0 and MDASI sadness severity, no significant differences were observed. For anxiety significant differences were observed using the PROMIS-29 anxiety profile v1.0 (ES = 0.30, p = 0.04) and the Spielberger State-Trait Anxiety Scale (ES = 0.39, p = 0.01). Though, Wyatt et al. [9] also used this scale but could not find any significant differences comparing reflexology, foot massage and a control group.

Eight other studies also did not find a significant difference comparing groups [8, 9, 21,22,23,24,25,26].

Wyatt et al. [25] conducted a sequential multiple assignment trial comparing reflexology to meditative practices, also including a control group. After 4 weeks nonresponding patients were randomized 1:1 to either the same group or the other group, while responsive patients continued their treatment for another four weeks. No significant results were reported for anxiety and depression. Tsay et al. [8] (n = 62, day 2–6 after surgery) did not find any significant differences for anxiety on average over all measurement points but a decrease in symptoms which was significantly stronger in the intervention group (β_I = -1.12 (0.49); p = 0.0231). This also applies to a study by Murat-Ringot et al. [21] (n = 80) in which reflexology was compared with a control group. Patients received four sessions of reflexology (30 min each) every two to three weeks during chemotherapy infusion depending on the chemotherapy protocol for four cycles. In a study by Rezaei et al. [24] (n = 74) patients received two sessions in one day but no significant differences were observed after the intervention compared to a control group. Sharp et al. [22] (n = 183, reflexology + SIS, scalp massage + SIS, self-initiated support for eight weeks), Ross et al. [23] (n = 26, reflexology, foot massage for six weeks), Sikorskii et al. [26] (n = 256, reflexology and control for four weeks) and Wyatt et al. [9] (n = 286, reflexology, foot massage, control for four weeks) did not find any significant results, as well.

Only one of the seven studies in which a certified reflexologist applied the intervention showed significant results in favor of the intervention [12] and only two out of five when it was applied by a caregiver [15, 16]. This is also the case for all three studies, in which no detailed information are given [17, 19, 27].

Methodical assessment of studies

In three studies [17, 19, 27], no information is given on other treatments, medication and comorbidities. Göral Türkcü et al. [27] also did not provide information on the control group but only on the intervention group after the second and final week of intervention which results in risk for reporting bias regarding the short term effect. Three other studies [12, 15, 16] display some methodical drawbacks. Stating limitations of their study, Sikorskii et al. [16] noted that there are methodological drawbacks so the results are not reliable for depression. Samancioglu Baglama et al. [15] did not test for normal distribution of data while the study design of the study by Stephenson et al. [12] doesn’t allow for interpretation of results. Murat-Ringot et al. [21] allowed home application of reflexology but did not consider it in their analysis. Consequently, not all data necessary for interpretation of the results is available. Sikorskii et al. [26] did not provide information on drop outs at all and Ross et al. [23] had a drop out of a third which was possibly caused by foot discomfort as this was noted as a common side effect. The latter, additionally, had a small sample size (n = 26) to begin with. In the study by Rezaei et al. [24], it is not clear who evaluated the data, so it cannot be ruled out that the reflexologist was involved here. Furthermore, the researcher stood at the patients’ bedside in the control group, with no information on possible verbal interaction between them. Due to the attention patients might have received hereby, this passive control group could possibly be considered as an active control. Wyatt et al. [25] randomized patients a second time depending on their outcomes after the first four weeks. Since this is not a complete randomization anymore and includes a high risk of bias, we only considered results of the first randomization. Furthermore, patients might have received varying frequencies of intervention making it hard to compare.

Fatigue

Description of studies

With fatigue being a very common side effect in cancer patients, nine studies investigated whether reflexology could be a useful tool in alleviating these symptoms. In four of the studies investigating the effects on fatigue reflexology was delivered by certified reflexologists [5, 9, 13, 28, 29] in three by caregivers [15, 16, 25] and in two it either isn’t clear who performed the intervention [30] or if the researcher applying it had any qualifications for doing so [18].

Four studies used an active control group [9, 13, 15, 18, 25] while the other five used a passive one [5, 16, 28,29,30].

The study by Rambod et al. [5], which included only patients with Hodgkin- and Non-Hodgkin Lymphoma, used the Multidimensional Fatigue Inventory and found significant differences between groups in favor of reflexology after the five day intervention period (p < 0.001). Significant differences after the intervention were also found for four of the five subdimensions of the inventory: general fatigue (p = 0.006), physical fatigue (p = 0.01), reduced activity (p = 0.01) and reduced motivation (p = 0.05). Nourmohammadi et al. [30] included 60 patients and obtained significant results in favor of reflexology between groups two months after the end of the four week intervention period (p = 0.000), showing possible long-term effects of the intervention. Conducting a pre-to-post comparison, Hesami et al. [28] included 80 subjects and, also using the Fatigue Severity Scale, found a significant difference between groups (p = 0.016) with less fatigue in the reflexology group. In the study by Dikmen et al. [18], the authors reported significant differences between groups at the 3^rd, 8^th and 12^th (follow up) week for both fatigue severity (3rd: p = 0.001; 8^th: p = 0.001; 12^th: p = 0.039) and effects of fatigue on daily life (all p-values = 0.001) with the lowest scores being reported in week eight for reflexology plus progressive muscle relaxation. Özdelikara et al. [29], who included 60 patients, observed significant differences between groups for fatigue severity (p < 0.05) and daily life activity exposure levels (p = 0.05) after the fourth chemotherapy treatment cycle. While investigating adverse effects, Uysal et al. [13] found that the reflexology group presented with significantly less grade 2 + fatigue when compared to the foot massage and control group in week 3 (p = 0.03), 4 (p < 0.001) and 5 (p = 0.036). Sikorskii et al. [16] used three different measurements to assess fatigue, only finding significant differences between groups using the MDASI fatigue severity scale (p = 0.03). Two other studies did not find any significant differences [9, 25]. Samancioglu Baglama et al. [15] found no significant differences during and after the intervention but already at baseline, showing more fatigue in the intervention group (p = 0.01).

Three of the four studies in which a certified reflexologist applied the intervention showed significant results in favor of the intervention [5, 28, 29], only one of the three in which a caregiver applied it [16] and both when no detailed information exist on this matter [18, 30].

Methodical assessment of studies

Four studies show some methodical problems [15, 16, 29, 30]. Sikorskii et al. [16] did not provide information on significance for baseline differences between groups whereas in the study by Nourmohammadi et al. [30], randomization was done based on days of the week leading to a high risk of bias. They also gathered information on patients’ believe in the impact of palliative practices showing that it was 20% higher in the reflexology than in the control group. This might have created a placebo effect since blinding factually couldn’t be achieved. Özdelikara et al. [29] did not control for multiple testing and there is also a potential risk for performance bias since patients were lying on ergonomic beds during reflexology sessions and there is no information on whether the control group was allowed to use these beds as well at some point. Baseline differences for fatigue in the study by Samancioglu Baglama et al. [15] put both the randomization and the validity of the result into question.

Quality of life / symptom severity and distress

Description of studies

In eight of the studies investigating Quality of Life or Symptom Severity and Distress the intervention was applied by a certified reflexologist [6, 9, 13, 21,22,23, 31, 32]. Three studies had caregivers apply the intervention [20, 25, 26] while in four studies no information are given on whether the researcher had any qualifications regarding reflexology [17, 18, 27, 33].

Eight studies used active control groups [9, 13, 17, 18, 22, 23, 25, 33], while the other six used passive groups [6, 20, 21, 26, 27, 31].

Quality of life, physical and social functioning and symptom distress and severity are all composed or representative of multiple symptoms, therefore showing a broader picture of the condition of the patients. Mantoudi et al. [17] examined the difference between a reflexology and a relaxation group for QoL over six weeks and found significant differences in change from baseline to after six weeks for the mental component summary score (p = 0.017, η² = 0.071) and the physical component summary score (p < 0.01, η² = 0.168). In a study by Göral Türkcü et al. [27] the global quality of life scale showed a significant difference between groups two weeks after the intervention [intervention: mean = 60.22 (SD = 17.17), control (mean = 40.59 (SD = 9.06), p < 0.01)] which was also found for the functional scale (p < 0.001) and symptom scale (p < 0.001). A study by Hodgson et al. [32], which included 12 people, compared reflexology with placebo reflexology on day one, three and five of their stay in the hospital. They reported a significant difference for the subcomponent of breathing (p = 0.026) and overall (p = 0.004). Dikmen et al. [18] also reported significant results for the 3^rd, 8^th and 12^th week (follow up) (p < 0.05), with the highest scores of quality of life being reported in the 8^th week [mean = 6.11 (SD = 0.274)] in the group receiving reflexology plus progressive muscle relaxation. No significant results for reflexology were found in a study by Sharp et al. [22] using FACT-B: TOI at 18 weeks post surgery but 24 weeks post surgery where reflexology plus self-initiated (SIS) support lead to a better outcome than SIS alone (p = 0.02) but did not show a significant difference when compared to the scalp massage plus SIS group. Using the Functional Assessment of Cancer Therapy-breast cancer version (FACT-B) total score a significant difference between the intervention and SIS group was detected at 24 weeks post surgery, as well (p = 0.03). A study by Kurt et al. [6] using the European Organization for the Research and Treatment of Cancer Quality of Life Questionnaire Chemotherapy-Induced Peripheral Neuropathy (EORTC QLQ-CIPN-20) only found a significant difference between intervention and control group in the last week of the six-week intervention period for the sensory function subscale (p = 0.024) while the other two subscales did not show any significant results. Uysal et al. [13] investigated adverse effects and found significantly less pain and fatigue (see above for detailed results). Furthermore, significantly lower grade 1 + urinary frequency in week 5 (p = 0.044) and grade + 1 distension in weeks 4 (p < 0.000) and 5 (p < 0.000) were found for the reflexology group compared to control. This study and two other studies [21, 31] also measured quality of life using the EORTC-Quality of Life Questionnaires (QLQ)-C30. Uysal et al. [13] found significantly better results on the function scale for the reflexology group compared to control for week 3 (p < 0.000) and 5 (p < 0.000). On the symptom scale significantly better results were found in the reflexology group compared to control (p = 0.003) while the reverse was found in week 5 (p < 0.000). Comparing groups for the global health scale, significant differences in favor of reflexology compared to both other groups were detected in week one (p = 0.012), three (p < 0.000) and five (p < 0.000). Özdelikara et al. [31] too examined the QoL and significant differences between groups for posttest measurements (24h after last chemotherapy cycle) for the general health score (p = 0.000), function score (p = 0.000) and symptom score (p = 0.000) were detected. Murat-Ringot et al. [21] did not find any significant differences. Wyatt et al. [9], using QoL FACT-B in their study also did not detect any diversity between the reflexology and the foot massage or control group when examining quality of life, as well as another study by Wyatt et al. [20], which used the Quality of Life Index and included 256 patients, where reflexology was compared with attention by the caregiver. Patients received at least one caregiver delivered reflexology session per week (real average 1.1) for the first four weeks. After that, there was no requirements and patients reported an average of 0.6 sessions per week until follow up in week 11. This study [20] also found significantly better results for the intervention group for summed symptom severity using MDASI and an adjusted coefficient of group variables over week five and eleven (p = 0.02) and significantly better results in week two (p < 0.01), 3 (p = 0.01) and five (p < 0.01) which can mainly be attributed to improvements in pain and fatigue. Sikorskii et al. [26], including 209 patients, also investigated symptom severity in an additional secondary analysis of this study by Wyatt et al. [20] only taking into account the first four weeks. Significantly better results in favor of the reflexology group were found for pain (p = 0.03) with no significant results in any of the other domains. Ross et al. [23] did not find any significant results looking at symptom distress except reportedly significantly greater appetite and mobility in the foot massage group, but no p-values were given. Dyer et al. [33] included 115 patients, who received four sessions of either aromatherapy or reflexology over the course of ten weeks on average. Results of the first concerns score of the Measure Yourself Concerns and Wellbeing (MYCaW) show a significant difference in favor of aromatherapy (p = 0.046) while the second concerns score shows no significant difference between groups but a significant improvement within groups (no p-values for comparison). This is also the case for overall wellbeing scores (no p-values for comparison). The study by Wyatt et al. [25] also found no significant differences between groups for symptom severity.

Four of the six studies in which a certified reflexologist applied the intervention showed significant results in favor of the intervention [6, 13, 22, 31], two of three when caregivers applied it [20, 26] and three of four when no detailed information exist [17, 18, 27].

Methodical assessment of studies

In the study by Uysal et al. [13], not only did the reflexology group receive longer sessions than the foot massage group but there were also significant differences in global QoL between groups at baseline. Furthermore, the control group was comprised of much more patients with grade III compared to grade II tumor than the reflexology group, which could possibly have influenced the patients’ general wellbeing. Finally, the authors did not provide information on results of EORTC QLQ CR29 as planned, resulting in a reporting bias. Wyatt at el. [20] investigated both symptom severity and QoL. However, since the authors did not describe what attention by the caregiver in the control group as an intervention looked like, it is hard to draw a deduction for the actual efficacy. In the study by Hodgson et al. [32], no consistent timing of the post intervention questionnaire existed in the beginning, according to the authors items were left out from the questionnaire due to printing errors and other information like homogeneity between group and p-values are missing. Dyer et al. [33] failed to present p-values for a group comparison for the secondary concern of the MYCaW score so no conclusion can be drawn from this outcome.

Physical and social functioning / interference with daily life

In one of the three studies examining these endpoints the intervention was applied by a certified reflexologist [9] while the other two had caregivers apply it [16, 20].

One study used an active control group [9] and two studies used a passive one [16, 20].

Wyatt et al. [9] investigated physical functioning and discovered significantly better results in the reflexology group compared to control (p = 0.04) but found no significant differences between reflexology and foot massage. In addition, the effect on dyspnea was measured showing significantly better results for reflexology when compared to control (p < 0.01) and foot massage (p = 0.02). In another study, Wyatt et al. [20] found no significant differences between groups for physical functioning and satisfaction with participation in social roles, while they observed significantly better results for reflexology using an adjusted coefficient of group variables over week five and eleven (p < 0.01) and significantly better results in week two (p = 0.02), 3 (p < 0.01) and 5 (p < 0.01). Sikorskii et al. [16] found no significant differences between reflexology and a control group when comparing different PROMIS and legacy measures for physical functioning. Comparing groups for social functioning, however, significant differences in favor of reflexology were found using SF-36 social functioning (legacy) (p = 0.04), while PROMIS-29 satisfaction with participation in social roles showed insignificant results.

The only study [9] where reflexology was applied by a certified specialist showed significant results in favor of the intervention for physical functioning but not interference with daily life while results are very mixed for the studies in which caregivers applied it. Sikorskii et al. [16] and Wyatt et al. [20] each showed significant results in favor of the intervention for only one of the above outcomes.

Methodical assessment of studies

In the study by Wyatt et al. [9], there is a risk for sampling bias and they did not report results comparing the two active groups except for dyspnea. In another study by Wyatt et al. [20], no information are provided regarding details on the control group, which received attention by their caregivers. Sikorskii et al. [16] did not provide information on the homogeneity of groups.

Nausea and vomiting

In all the studies investigating nausea and vomiting reflexology was applied by a certified reflexologist.

Only one study used an active control group [9] while the other three studies used a passive one [10, 21, 29].

Two studies investigated the effect of reflexology on chemotherapy induced nausea and vomiting at which only the study by Özdelikara et al. [29] found significant differences between groups using the Rhodes index of nausea, vomiting and retching which is composed of three subscales. A significant advantage for reflexology was found for the subscale of symptom development scale (p = 0.000) and distress development (p = 0.000) after the 4^th cycle of CTX, while none was found regarding if they experienced symptoms or not. Murat-Ringot et al. [21] measured CINV during the second cycle of chemotherapy, asking patients to fill out a VAS before and after the reflexology intervention or upon entering and leaving the hospital for the control group respectively. An intention-to-treat analysis was conducted, with patients with missing outcome data being considered as having an increase of > 2 on the VAS. While the per-protocol analysis showed an advantage for the intervention (p = 0.001), the intention-to treat did not find an effect. Two more studies also investigated nausea only, with Anderson et al. [10] (n = 40; one session with pre and post test), who used VAS and only looked at the in-group difference, and Wyatt et al. [9] who used the nausea item from the physical subscale of FACT-B, both not finding significant results.

Methodical assessment of studies

As mentioned above, the study by Wyatt et al. [9], shows risk for sampling bias and incomplete reporting. Murat-Ringot et al. [21] allowed home application of reflexology but did not consider it in their analysis. Therefore, data is missing for interpretation of the results in its entirety. In the study by Özdelikara et al. [29], it was not controlled for multiple testing and there might be a risk for performance bias, whereas Anderson et al. [10] did not provide information on homogeneity between groups.

Sleep, mood and relaxation

In three studies the intervention was carried out by certified reflexologists [5, 7, 22], in one study caregivers applied it [16] and in one there are no information on the researchers qualifications who applied it [33].

Three studies used an active control group [7, 22, 33], while two used a passive control group [5, 16].

Rambod et al. [5] found significant differences between groups in favor of the intervention group for two of the scales when investigating sleep quality after five days of reflexology (subjective sleep quality, p < 0.001; sleep latency, p = 0.001). Total sleep quality at baseline already showed better values for the reflexology group (p = 0.05), though, suggesting a potential problem with randomization. After the intervention differences were still significant (p < 0.001). No significant differences were found in this study by Sikorskii et al. [16] comparing sleep disturbance using PROMIS-29 and MDASI (legacy). Hodgson et al. [7] compared groups for affect by observing patients four times per day on intervention day for four weeks for five minutes each time and then averaged measures for mean values. No significant differences between groups for both negative and positive mood were found. Sharp et al. [22] found significant differences between groups at 18 weeks after surgery in favor of reflexology (p < 0.0005) and scalp massage (p < 0.0005) compared to control and significant differences for reflexology compared to control at 24 weeks post surgery (p = 0.02) using the Mood Rating Scale (MRS) relaxation subscale. The easy-goingness subscale also revealed significant differences in favor of reflexology compared to scalp massage (p = 0.04) and control (p < 0.0005) at 18 weeks post surgery. Dyer et al. [33] compared pre and post session scores for relaxation between reflexology and aromatherapy groups for all four sessions and for change over all four sessions which on average were distributed over ten weeks but did not find significant differences for both.

Two of the three studies where reflexology was applied by a certified specialist showed significant results in favor of the intervention [5, 22], while no such benefits could be observed for all other studies.

Methodical assessment of studies

As mentioned above, the statistical analysis in the study by Hodgson et al. [7] is incomprehensible and no information was provided on homogeneity of groups [16].

Narcotic analgesia consumption

In the study by Tsay et al. [8], the intervention was applied by a certified reflexologist a passive control group was used. It is the only one included which also investigated the influence of reflexology on narcotic analgesia consumption as an outcome providing reflexology on days two to four after cancer surgery. At follow up on day five and six after surgery for hepatocellular or gastric carcinoma the intervention group showed a significantly lower use in Demerol than the control group (p = 0.015). However, there might be an interference of analgesics with the intervention.

Self Esteem / psychiatric morbidity

In both studies listed here the intervention was applied by a certified reflexologist.

One study used an active control group [22] and one used a passive one [21].

Murat-Ringot et al. [21] measured self-esteem and found no significant differences between groups at the end of the study. At the end of the study a Body Image Questionnaire was used while baseline values were collected using the Rosenberg self-esteem scale, which makes interpretation over the course of the study difficult. Furthermore, as mentioned above, home application of reflexology was not accounted for in the analysis. Sharp et al. [22] investigated psychiatric morbidity and found no significant differences between groups.

Perceived social support / quality of relationship between caregiver and patient

The intervention was applied by caregivers in this study, which used a passive control group.

No significant differences between groups were found by Wyatt et al. [20] looking at perceived social support and the quality of the relationship between caregiver and patients. However, as noted previously, no information are provided regarding details on what the intervention in the control group looked like, who received attention by their caregivers.

Adverse events

No adverse events that can be attributed to reflexology were reported.

Discussion

An overall problem in designing studies with an active intervention is that true blinding of patients is very hard to achieve, since patients are aware of the application of an intervention. A possibility to blind a patient is by applying a very similar technique to the same body part as done by a very small number of studies included. Still, there is no way to blind the people applying the intervention and/ or the control counterpart. Therefore, while often termed as single or even double blind, most studies included have an open design.

As pain is arguably one of the most relevant side effects cancer patients experience, 13 of the included studies investigated the effect of reflexology on pain. Nine of the studies showed at least partially significant results [5, 8, 11, 12, 14,15,16, 18, 19]. The study by Dikmen et al. [18] found that reflexology has a positive effect on pain. However, some strong methodical drawbacks such as incomplete reporting of information should be considered and the results should be viewed with caution. Other studies also lack information on homogeneity of groups [11, 16] or dropouts [19] whilst one study also shows risk for a reporting bias [14]. Incorrect analysis of the study by Stephenson et al. [12] doesn’t allow for interpretation of the results. Two of the studies showing significant advantages for reflexology [5, 11] only investigated the effect over a very short time (five days; directly and after 3h, 24h respectively), which might indicate an acute effect on pain. While five other studies’ results where insignificant [6, 7, 9, 10, 17] they also presented with similar problems.

All in all, due to some strong methodical drawbacks these findings should be viewed with caution and a clear conclusion cannot be deducted. As none of the studies investigating pain allows for true blinding, it is unclear if the positive effect is attributed to the intervention or a result of being relaxed, as the relaxation response might help alter pain perception [34].

As cancer patients get confronted with their diagnosis and the consequences, dealing with potential anxiety and depression is important. Seven studies reported significant effects of reflexology on anxiety and depression [12, 14,15,16,17, 19, 27]. Mantoudi et al. [17] only found a significant advantage for depression but not for anxiety after four and six weeks. More information on other treatments, medication and comorbidities would have been of value here due to the possibly multifactorial origin of anxiety and depression but no further information is given. The same lack of information also applies to Jahani et al. [19] and Göral Türkcü et al. [27] while the latter also show risk for reporting bias regarding the short term effect. Three other studies [13, 16{Sikorskii, 2018 #344]} showed significant results for anxiety. However, the study design in the study by Samancioglu Baglama et al. [15] doesn’t allow for data interpretation, while the study by Sikorskii et al. [16], which also showed partially significant results for depression, presented with measurement and statistical hypothesis testing errors and therefore unclear results.

Eight other studies [8, 9, 21,22,23,24,25,26] reported insignificant results. Murat-Ringot et al. [21] allowed home application of reflexology but did not consider it in their analysis, whereas other studies showed some methodical problems [23, 24, 26]. The study by Wyatt et al. [9] did not blind patients even though it would have been possible since the active control group received a foot massage.

Overall, these findings described on anxiety and depression do not hint to a benefit by reflexology. Studies that reported significant results in favor of reflexology are presented with some major drawbacks. Additionally, there was only one study which allowed for true blinding of patients [23]. The evidence supporting long term effects is very thin as only two of the studies [19, 27] investigating these effects showed significant improvements. This leaves the impression that reflexology can at best help in improving anxiety and depression in an acute scenario. This could possibly be explained by an increased level of relaxation, which wears off after returning to usual life with all its stressors.

Taking a closer look at the effects on fatigue, six out of nine studies showed significant results in favor of reflexology [5, 16, 18, 28,29,30]. Three of them show some methodical problems [16, 29, 30], such as lack of information on baseline comparisons [16], risk for placebo effect [30] and no control for multiple testing [29]. The results of two other studies [5, 28] indicate that reflexology might be a tool to mitigate chemotherapy-induced fatigue, especially physical fatigue, in the short term.

Three other studies [9, 15, 25] did not find significant results. In the study by Wyatt et al. [25], patients might have received varying frequencies of intervention making it hard to compare, while in the study by Samancioglu Baglama et al. [15], the control group already showed significantly less fatigue than the reflexology group at baseline. One might also argue that baseline differences in fatigue could also influence the perception of the two other parameters (pain and anxiety) examined in this study, which while showing significantly better results in the reflexology group, nonetheless, could have possibly shown even stronger effects.

To conclude, the above shortcomings need to be considered. Although the trend indicates that reflexology might have a positive impact on fatigue in cancer patients it remains unclear whether this also pertains to long term effects as only one of the above studies investigated possible long term effects but did not find significant results [9]. As fatigue in the context of cancer is a chronic state it is important to note that short term effects only offer a brief symptom relief. This implies that frequent application would be required for a meaningful impact. Furthermore, none of the studies investigating fatigue allow for true blinding of the patients. Since all the studies showing significant results except one [18] only had a passive control group, it is unclear whether other factors beyond the intervention might have played a role in the short term improvement.

Quality of Life is arguably the most comprehensive measure for cancer patients as it is more so a combined than a single outcome and therefore allows for covering more influencing factors.

10 studies presented with significant results in favor of reflexology [6, 13, 17, 18, 20, 22, 26, 27, 31, 32]. While the study by Mantoudi et al. [17] indicates that patients experience less restrictions of daily life due to physical limitations when using reflexology rather than simple relaxation, the study by Kurt et al. [6] suggests very little effect, as only a small part of the results was significant. Sharp et al. [22] found significantly better results compared to the control but not to the scalp massage group which implies that more than one form of physical intervention might result in the desired outcome. Three other studies come with some drawbacks [13, 20, 32] like differences in session length and risk for a reporting bias [13], lack of information on what attention by caregivers looked like [20] and missing information on p-values and more [31]. For this reason, the results of these three studies should be viewed with caution.

Five other studies did not produce significant results [9, 21, 23, 25, 33] and two of them present with methodical drawbacks [25, 33]. Dyer et al. [33] reported statistically and clinically relevant intragroup improvements for the aromatherapy and reflexology group but failed to present p-values for a group comparison for the secondary concern score. Due to this, no conclusion can be drawn from this outcome.

Overall, the results regarding QoL and symptom distress and severity are mixed. While more studies speak for a positive effect, some drawbacks limit the informative value. This again includes a lack of true blinding, which applies to only one study [23]. Therefore, the influence of a placebo effect should at least be considered.

The way patients can continue to navigate daily life is closely related to their QoL. In the study by Wyatt et al. [9], significant results in favor of reflexology were observed for physical functioning when comparing reflexology to a control group but not when compared to foot massage. Even though there was no blinding, this shows that reflexology is likely not better than other similar interventions for this outcome. However, reflexology might help with dyspnoe as examined in this study, as well. With Sikorskii et al. [16] showing a positive tendency for social but not physical functioning and Wyatt et al. [20] showing a significant improvement for interference with daily life but not physical functioning, the overall results are pretty ambiguous and don’t allow for a clear trend.

As it is a common side effect of chemotherapy, four studies [9, 10, 21, 29] also investigated whether reflexology might be a useful tool in alleviating nausea and vomiting. Three of them [10, 21, 29] investigated short term effects but only one [29] found at least partially significant results in favor of reflexology. However, Murat-Ringot et al. [21] showed that reflexology might potentially help reduce the dosage of antiemetic drugs needed to deal with delayed nausea and vomiting. Altogether, the trend points towards no significant efficacy of reflexology on nausea and vomiting, though.

Only a handful of studies reported on sleep, mood and relaxation [5, 7, 16, 22, 33]. Even though Sharp et al. [22] and Rambod et al. [5] found a significant effect, the overall results indicate that reflexology doesn’t seem to be superior to other interventions with the same goal.

While speaking for possible positive effects on the consumption of narcotic analgesics, only one study [8] investigated this outcome. Therefore, the evidence is too limited for a conclusion. This also applies to the outcomes of self esteem [21], psychiatric morbidity [22], perceived social support and quality of relationship between caregiver and patient [20], which all presented with insignificant results. Consequently, reflexology presumably is not an effective tool here.

Since no trend could be observed in terms of who applied the intervention, it likely does not make a difference. This once again raises the question, whether it is the intervention itself or simply the psychological and physical attention received by the patients, that has led to some significant benefits in favor of reflexology.

The studies included in this review investigated a variety of symptoms which we discussed mostly individually. However, it is important to note that changes in characteristics of one symptom and changes in the current state of disease might influence one another, as proposed in the biopsychosocial model [35]. Since no individual data on patients exists in the included studies, exploring such interactions for the most part is beyond this review.

Limitations of this work

This review has a few limitations. All studies exclusively included adults which doesn’t allow for conclusions regarding children. Furthermore, only studies in German or English as well as only Randomized Controlled Trials were included, excluding grey literature. Something else to consider is that most studies show a high risk of bias with a small number showing a moderate risk. Additionally, some outcomes were only investigated by a small number of studies.

Conclusion

Studies on reflexology included a wide variety of different types of cancer not restricting conclusions to a small group of cancer types. The reported results are very heterogenous. Most studies indicate that reflexology is superior to a passive control group for pain, quality of life and fatigue but not anxiety and depression. For other outcomes, the sample of studies is too small for a conclusion. As results are very mixed, no trend in efficacy could be observed looking at whether reflexology was performed by a certified professional or a caregiver. The methodical quality of the majority of studies is too poor for them to demonstrate proof for the specific efficacy of reflexology. Meanwhile, it appears that reflexology is not superior to other massage interventions as there exists no physiological concept on how these reflex zones work. Reflexology rather seems to draw its efficacy from the care and attention received through the intervention. This, in fact, can be achieved by any form of massage.

For future randomized controlled trials on reflexology in oncological treatment we would therefore like to recommend a few criteria to avoid possible bias. Control groups should be active with an intervention that’s as indistinguishable as possible from reflexology for the patient such as foot massage. This would allow for real blinding of patients. While a bit more extensive, possible subanalyses of patients who believed in such interventions prior to the trial and patients who did not could help shed more light on possible influencing factors. This could also be applied for other factors that are considered part of the biopsychosocial model, such as stress unrelated to the disease and patients’ support networks. Additionally, a protocol where all patients receive the intervention in the same time interval regarding their cancer therapy might be beneficial. This could help reduce the impact differences in time intervals between reflexology and cancer treatment might have on symptoms.