Introduction

For over 100 years, physicians have attempted to treat cancer patients by heating tumour tissue. While certain techniques allow to directly kill cancer cells by heat, for whole-body and electro hyperthermia, the rationale is that an increase in the temperature of the tumour (40–43 °C) [1] induces changes in perfusion and re-oxygenation, produces heat-shock proteins, stimulates immunological activity and thus improves tumour response to radio- and chemotherapy [2, 3]. Previous researches in conventional medicine concluded that heat and radiation may have synergistic effects [4], because cells in late S phase are resistant to radiation and studies found that they were additionally most sensitive to heat in that phase [5]. In addition, tumour tissue is characterized by nutrient deprivation, limited oxygenation and highly acid conditions [6]. These properties are connected with low effects on tumour response to radiation [7]. Temperatures between 40 and 42 °C may increase perfusion and thereby oxygenation increases [8] with the possibility to overcome hypoxia-associated radio-resistance [9]. On the one hand, the higher sensitivity to chemotherapy is a result of elevated tumour blood flow, resulting in higher therapeutic drug concentrations in the tumour tissue [10]. A clear temperature-dependent increase in drug uptake has been shown in preclinical trials for example for cisplatin [11], carboplatin [12] or Melphalan [10, 13]. On the other hand, hyperthermia is able to increase the number of DNA strand breaks induced by chemotherapy [10, 12, 14] and may inhibit DNA repair [15]. Moreover, induction of heat-shock proteins (HSP) and immunomodulation has been described [16].

Nowadays, different methods of hyperthermia may be distinguished [1]. Well-established usage is hyperthermic intraperitoneal chemotherapy (HIPEC) [17], the combination with chemotherapy in patients with sarcoma [18, 19] or the combination with radiotherapy for patients with cervical cancer who refuse or are not eligible for chemotherapy [20, 21]. Moreover, interstitial hyperthermia with brachytherapy is used for local applications [22]. Other methods like the activation of magnetic nanoparticles by an alternating magnetic field are currently being explored as a technique for targeted therapeutic heating of tumours [23]. Thermal ablation, which uses temperatures of above 100 °C to destroy tumour tissue directly, has to be differentiated from these types of hyperthermia treatment [24]. For hyperthermia treatments, the European Society of Hyperthermic Oncology (ESHO) and the Interdisziplinäre Arbeitsgruppe Hyperthermie (IAH) have defined quality standards [25,26,27,28,29,30]. A decisive criterion is that the temperature is measured directly in the tumour in order to verify heating of the target volume to the required 40–43 °C. This is an important point to differentiate between conventional and alternative methods, since no intra-tumoural temperature measurement is performed in the alternative hyperthermia methods.

Electro hyperthermia (EH) does not comply with the standards mentioned above and is widely spread. The method of capacitive coupling seems to be particularly popular because tumour cells may allegedly be heated selectively with this method. The therapy takes advantage of the supposedly special absorption rate of the extracellular fluid of the tumour. Tumour tissue is said to have a lower impedance than neighbouring tissue, so that energy is absorbed primarily by the tumour. This is supposed to achieve self-focusing [31]. A well-known representative of capacitive coupling is oncothermia [32]. This method uses radiofrequency waves with a frequency of 13.56 MHz [33].

Whole-body hyperthermia (WBH) has been evaluated in the 1950s to 1980s, but was left due to a negative risk/benefit ratio. So-called moderate whole-body hyperthermia is still used [34]. In alternative medicine, hyperthermia is used as WBH with moderate temperatures (about 39 °C) or high temperatures (more than 40 °C) [31]. Heating is reached by a warmed waterbed or by infrared radiation. Other alternative WBH procedures are the extracorporeal heating of blood or the induction of fever by bacterial toxins [3].

Therefore, we have included in this review only hyperthermia methods that do not belong to conventional medicine and titled these alternative methods. This includes all methods that heat the whole body either from the outside or from inside by using extracorporeal circulation or bacterial toxins. In addition, all hyperthermia methods with electric fields that do not meet the ESHO criteria are considered. Conventional other medical procedures with heat such as HIPEC [35], HIVEC [36] or thermoablation [24] are not part of our review. Moreover, hyperthermia generated by high-frequency radiofrequency waves or with microwaves [37] with an adequate real-time thermal dose monitoring [38, 39] meeting the requirements of the ESHO is also explicitly not part of our review [25,26,27,28,29,30].

The aim of our systematic review is to assess the evidence of these different alternative methods of external hyperthermia provided beyond the international standards.

Method

Inclusion and exclusion criteria

Inclusion and exclusion criteria are listed in Table 1 based on a PICO model. Generally, all study types were included if they reported patient-relevant outcomes after the treatment of adult cancer patients with a complementary medical hyperthermia. Complementary hyperthermia methods included any modalities whether whole-body or electro hyperthermia, because exactly these do not meet the defined quality criteria of the European Society of Hyperthermic Oncology [25,26,27,28,29,30]. Any kind of comparison was eligible in this review. This includes watch and wait, standard care, sham and placebo. Because of the wide range of application fields, all cancer entities were included. Since little high-quality evidence was expected, systematic reviews and randomized controlled trials were included as well as controlled trials, one-armed studies, case reports and case series. Criteria for rejecting studies were primary prevention, grey literature, other publication types than primary investigation/reports (e.g. comments, letters, abstracts) and study population with more than 20% children (patients under the age of 18) if results of adult patients with cancer were not reported separately or precancerous conditions. Additionally, systematic reviews, randomized controlled trials and cohort studies were excluded if they reported only not patient-centred outcomes at all (only labour parameters). Furthermore, we included single-arm studies, case reports and case series if side effects were reported. Language restrictions were made to English and German. Studies that evaluated a combination of hyperthermia and other treatments versus none of the parts of the combination were not included as it would not be possible to determine the impact of hyperthermia. Nevertheless, we analysed the side effects of hyperthermia treatment in these 43 studies and reported the side effects, which were clearly attributed to hyperthermia treatment.

Table 1 Inclusion and exclusion criteria
Full size table

Study selection

A systematic review was conducted using five databases (Medline (Ovid), CINAHL (EBSCO), EMBASE (Ovid), Cochrane CENTRAL and PsycINFO (EBSCO)) in April to May 2021. 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 the different types of alternative hyperthermia therapy (Table 2). The search string was highly sensitive, since it was not restricted by filters of study or publication type. After importing the search results into EndNote X8, all duplicates were removed and a title–abstract screening was carried out by two independent reviewers (CL and SK). In case of disagreement, consensus was made by discussion. After that, all full texts were retrieved and screened again independently by both reviewers. When title and abstract did not have sufficient information for screening purposes, a full-text copy was retrieved as well. Additionally, bibliography lists of all retrieved articles were searched for relevant studies.

Table 2 Search strategies for each database
Full size table

Assessment of risk of bias and methodological quality

All characteristics were assessed by two independent reviewers (CL and SK). 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 analysed with the AMSTAR-Checklist Version 2.0 for the SR [40], the SIGN-Checklist for controlled trials Version 2.0 [41], the SIGN-Checklist for cohort studies Version 3.0 [42] and the IHE-Checklist for single-arm studies and case series [43]. In addition, blinding of researchers, blinding of outcome assessment and comparability of groups before treatment, not only in terms of demographic variables but also concerning the outcomes, were examined.

The included studies were rated with the Oxford criteria. Additional criteria concerning methodology were size of population, application of power analysis, dealing with missing data and drop-out (report of drop-out reasons, application of intention-to-treat-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 (CL) and controlled by two independent reviewers (JD, JH). As a template for data extraction, the evidence tables from the national Guideline of Complementary and Alternative Medicine in Oncological Patients of the German Guideline Program in Oncology [44] were used.

Results

The systematic research revealed 47,388 results. Eighteen studies were added by hand search. At first, duplicates were removed leaving 31,200 studies. Of these, 30,334 studies were rejected due to several reasons. After title–abstract screening 284 studies remained for full-text copy from which 231 studies were excluded due to following reasons: 16 studies did not use hyperthermia, another 125 studies did not use alternative hyperthermia, and in 43 studies, multiple interventions were administered at the same time so that an assessment of hyperthermia was not possible. Moreover, five studies were not published in English or German. For 11 abstracts, full text was not available, and four studies included only preclinical aspects. An inadequate article type (e.g. poster, letter to the editor, conference articles) was another reason for the exclusion of 27 studies. The flowchart of studies through the review can be seen in Fig. 1. A list of the studies excluded after full-text screening and the reasons for exclusion are presented in supplementary Table 3.

Fig. 1
figure 1

Consort diagram

Full size image
Table 3 Efficacy and side effects of hyperthermia in studies of the first level of evidence (SR, RCTs, CTs and cohort studies)
Full size table

Finally, 96 publications were analysed in this review: 53 studies on alternative hyperthermia and 43 studies including multiple interventions which were only considered with respect to side effects of hyperthermia. Detailed characterization of the included studies may be seen in Figs. 2 and 3 and in Tables 3, 4, 5, and 6. In the 43 studies with multiple interventions, only the side effects were analysed. The characteristics of these studies can be seen in Table 7, the relevant adverse events in Table 8.

Fig. 2
figure 2

Risk of bias of studies of the first level of evidence (short characteristics)

Full size image
Fig. 3
figure 3

Risk of bias of studies of the second level of evidence (short characteristics)

Full size image
Table 4 Efficacy of hyperthermia in studies of the second level of evidence (single-arm studies and case series)
Full size table
Table 5 Efficacy of hyperthermia in case reports
Full size table
Table 6 Adverse events in studies of the second level of evidence (single-arm studies, case series and case reports)
Full size table
Table 7 Characteristics of studies with multiple intervention
Full size table
Table 8 Adverse events related to hyperthermia treatment of studies with multiple interventions
Full size table

Characteristics of included studies

The first level of evidence included 14 studies, with one systematic review (SR), six randomized controlled trials (RCT), three controlled trials (CT) and four cohort studies with overall 1366 patients, from which 1137 were analysed, due to 229 drop-outs. The age of patients ranged from 17 to 86 years. In total, 955 participants were females, and 411 were males. In one study with 131 participants [50] data on age, gender and drop-outs were missing, whereas in the SR [45] data on age, gender, number of participants and drop-outs were missing. The publications by Minnaar et al. [51,52,53] used the same patient collective, but the endpoints were different, so the studies were analysed individually.

The second level of evidence consisted of 32 single-arm studies, one case series and six case reports comprising 712 patients. Of these, 332 were females and 363 were males. In one study, data on gender were missing for 17 patients [63]. The age of patients ranged from 12 to 86 years. The publications by Bruns et al. [61] and Bakhshandeh et al. [60] reported the same 27 patients. Because the focus was partially on different outcomes, these studies were analysed individually.

Overall, WBH was used in 33 studies (WBH in 24 studies [45,46,47,48, 59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74, 88, 92, 93, 96], ECC-WBH (WBH with extracorporeal circulation) in 7 studies [75,76,77,78, 89, 91, 97], in one study WBH and ECC-WBH each used for different groups [49] and in one study WBH (water-filtered infrared-A radiation) was administered together with therapeutic fevers [50]). EH was used in 20 studies [51,52,53,54,55,56,57,58, 79,80,81,82,83,84,85,86,87, 90, 94, 95]. Outcomes included tumour response, course of tumour markers, survival data, pain, quality of life or quality of recovery, body weight, ascites and fatigue. Further outcomes included haemodynamic, haematologic, serum chemistries, immunological and pH values, PaCO2 and respiratory rate. Other specific results contained measured temperatures, tumour oxygenation and heat-dose tolerances.

Risk of Bias in included studies

The main shortcomings of the studies are summarized in Fig. 2 (first level of evidence) and Fig. 3 (second level of evidence). Among the first level of evidence seven studies have an acceptable quality, six studies have a low quality, and one study is rated as unacceptable. The single-arm studies showed a quality from low to acceptable. A detailed description of the methodological quality of each study is provided in supplementary table 1 (first level of evidence) and supplementary table 2 (second level of evidence).

Endpoints in the first level of evidence (SR, RCTs, CTs and cohort studies)

Detailed information to the outcomes and side effects can be seen in Table 3.

Benefit and risks of hyperthermia in combination with systemic therapy on solid tumours (Whole-body hyperthermia)

In the SR by Lassche et al. [45] pre-treated patients with heterogeneous solid tumours were treated with WBH and CTx (chemotherapy). Among these, ovarian cancer (n = 3 studies), colorectal carcinoma (n = 2 studies), lung cancer (n = 2 studies) and sarcoma (n = 3 studies) were the most common entities analysed in this SR. In total, 13 studies included patients with various malignancies. These studies were then referred to as phase-I studies. CTx was used in all studies. These regimes were not uniform, but in most cases, platinum containing agents (n = 10 studies) or alkylating agents (n = 6 studies) were used. In the vast majority temperature reached 41.8 °C over a period of one hour and was generated by means of radiant heat. The response rate (including complete and partial response) was analysed in a total of 13 phase-I (various malignancies) and 14 phase-II (special malignancies) studies and varied between 12 and 89%. However, in the two studies with the highest response rates (86%, resp. 89%), the number of subjects was very small (n = 15, n = 18 resp.) and the pretreatment was described insufficiently. All the studies were single-arm studies. Grade 3 and grade 4 toxicities according to the CTAE criteria (Common Terminology Criteria for Adverse Events) occurred in almost all studies.

Recovery after surgery (Whole-body hyperthermia)

The RCT by Sulyok et al. [46] analysed the effect of WBH on the quality of recovery after curative colorectal cancer surgery (n = 18). The intervention group was sedated 3.5 h before the beginning of surgery and then treated with WBH by infrared radiation. The time at the target temperature of 39 °C was 2 h. Twenty-four hours after surgery, no significant difference was found for quality of recovery (QoR) overall or in quality of life (quality of recovery, global QoR-40, p = 0.81).

Effect of whole-body hyperthermia on the toxicity of chemotherapy (Whole-body hyperthermia)

The phase-II-CT by Hegewisch-Becker et al. [48] included 44 patients with adenocarcinoma of colon or rectum with progressive disease. Every second cycle of the biweekly chemotherapy regime, consisting of oxaliplatin, leucovorin and 5-fluoruracil, was combined with two hours of WBH with an estimated intra-tumoural temperature of 41.8 °C. WHO (World Health Organization)-grade three side effects of chemotherapy, including haematologic, gastrointestinal, peripheral neurological toxicities and fatigue, were rare and evenly balanced between the cycles with or without hyperthermia. Due to the crossover design, the comparability of the two groups is given, but carry-over effects cannot be ruled out.

Tumour response, abscopal response, survival and quality of life in cervical cancer patients (electro hyperthermia)

In the RCT by Minnaar et al. [51] 210 patients received RTx (external beam radiation) and cisplatin with or without EH. Heating reached estimated 42.5 °C and was generated via a capacitive heating device (EHY2000, Oncotherm) for two times a week. According to the authors, there was no need to measure the temperature, and the applied dose of EH should be controlled by means of the absorbed energy. Six-month local disease-free survival in the intervention group was significantly higher than in the control group (p = 0.003). Furthermore, a significant higher rate of complete metabolic responses was reported in the arm with the administration of EH (p = 0.005). A complete metabolic response as well in the primary tumour as in the lymph nodes within and outside the radiation field occurred significantly more often in the EH group (p = 0.013) [53]. The authors discussed that as abscopal effect which has been described in a few case reports in the literature due to radiotherapy. The quality of life (QoL) was determined using the EORTC questionnaires QLQ C30 and Cx24 6 weeks and 3 months post-treatment. Six weeks after treatment, mean change in cognitive function in the intervention group was significantly higher than in the control group (p = 0.031). Three months after treatment a significant improvement in social functioning (p = 0.049), emotional functioning (p = 0.017), fatigue (p = 0.037) and pain (p = 0.007) was reported in the EH group, compared to the control group [52]. No data on the target temperature in the tumour field are reported. In these studies, many calculations are performed. However, in the exact comparison of the intervention and control group regarding the therapy, these data are missing. Therefore, it is not possible to accurately compare the treatments between the two arms with and without hyperthermia. In addition, information about prior treatments is not specified and a description of possible additional co-interventions is missing. For the endpoints tumour response and local disease control, reasons for the drop-out of part of the participants are not given. Therefore, it cannot be excluded that for these endpoints only suitable patients were considered.

Tumour response and survival in breast cancer patients (electro hyperthermia)

In the RCT by Loboda et al. [54] 103 patients were treated with EH together with neoadjuvant CTx. The control group, containing 97 patients, received only the neoadjuvant CTx. Target temperatures were estimated to reached up to 38.8 °C and maintained for about 30 min. The intra-tumour temperature estimations were based on an unvalidated bioheat model combined with skin temperatures measured using a thermal imaging camera. Considering the individual values for the tumour response according to the RECIST criteria (Response Evaluation Criteria in Solid Tumours), no significant difference between the two groups was described. The 10-year OS in the hyperthermia group was significantly longer (p < 0.009). Some points must be questioned in this study. The abstract reports that the intervention group had achieved a significantly higher rate of objective responses. An objective response included patients with a partial and complete response. However, when looking at the individual values for a complete response, partial response or stable disease, no significant difference can be found. Therefore, the significant difference in the abstract has to be doubted. Moreover, no information about the method of randomization is indicated. Another point for statistical criticism concerns the 10-year OS. No exact figures are given, for example, by how much longer survival was in the intervention group. The calculation of a significant difference is thus not comprehensible and remains vague. In addition, a comparison of the altered blood flow after treatment between the two groups is lacking.

Survival and tumour response in patients with glioblastoma (electro hyperthermia)

In the CT by Mahdavi et al. [55] glioblastoma patients were treated with CTx and RTx. The patients in the intervention arm received additional EH with estimated 41 °C (Celsius 42 +) two times a week for one hour each. Although tumour volume was significantly lower after treatment in the intervention group (p < 0.05), the OS between the two arms after 18 months was not significantly different (p = 0.55). Furthermore, the Karnofsky performance status did not differ significantly. In total, no information about the allocation method is given. Moreover, no comparison of the patient and tumour characteristics at baseline was conducted at all and in addition with the small number of participants (n = 38) an unequal distribution of subjects cannot be excluded. Additionally, side effects were reported only briefly.

Pathologic outcome and survival in patients with rectum carcinoma (electro hyperthermia)

In the retrospective cohort study by Kim et al. [58], 120 patients with rectum carcinoma received neoadjuvant RTx and CTx. The intervention group was treated additionally with EH (EHY 2000, Oncotherm). A hyperthermia session lasted one hour, and the number of sessions varied between one and twelve. Information about the targeted temperature or the temperature measurement is not given. A comparison of the two groups was done at baseline. Regarding the pathologic outcome (near total regression and total regression), no significant differences between arms were found, except for tumours with an initial primary volume more than 65 ml (p = 0.024). Considering the 2-year OS (p = 0.73), the 2-year disease-free survival (p = 0.054), the 2-year locoregional recurrence-free survival (p = 0.09) and the 2-year distant metastases-free survival (p = 0.083) no significant difference was found between the two groups. It has to be noticed that the dose of RTx was different between the treatment groups. In the intervention group the RTx-dose varied between 40 and 50.4 Gy, and in the control group, all participants received 50.4 Gy. Furthermore, it is not clear how the allocation to the different treatment groups was made.

Tumour response and survival of glioblastoma and astrocytoma patients (electro hyperthermia)

In the retrospective cohort study by Fiorentini et al. [56] 111 glioblastoma multiforme and 38 astrocytoma patients were divided into two groups. The intervention group (n = 52) was treated with EH (EHY2000, Oncotherm), whereas the control group (n = 97) received best supportive care together with CTx. The temperature in the EH group was between 40 and 42.5 °C and was only estimated. In the subgroup of astrocytoma patients, the results show an overall positive response, including complete response, partial response and stable disease in favour of the intervention (p < 0.005). The same applies for the subgroup of glioblastoma patients (p < 0.05). Moreover, the median OS was significantly better in both subgroups compared to the controls (p = 0.0065; p = 0.047). In the control arm, different CTx-regimes were administered; in contrast, the intervention arm received no CTx at all. With regard to the methodology of the study, there are several drawbacks. The functional recovery was only measured by ECOG grading. A comparison of the baseline data between control group and intervention group is missing. Therefore, it is not clear whether both arms can be compared at all.

Pain relief in lung cancer (electro hyperthermia)

In this retrospective cohort study by Kim et al. [57], the intervention group was treated additionally with EH (EHY2000, Oncotherm) two to three times (estimated 39–42 °C for about 60 min). Data collection was carried out at four different time points during 180 days after the start of the study. No significant differences could be found at any time for pain intensity and effective analgesic score (EAS), while the changes of the EAS over the time distinguished in dependence of the treatment (p = 0.038) with worse values in the intervention group in the first 60 days (p = 0.030). Exactly in these days, a significantly higher opioid analgesic dose was used in the intervention group (p = 0.022).

It must be critically noted that at baseline only 47.4% of the initially matched control group entered the study. Owing to this high drop-out, the comparability is limited and perhaps not given anymore. Despite that, the results suggest that EH leads to more pain immediately after the treatment.

Effect of preceding WBH on induced therapeutic fevers

In the retrospective phase-I-CT by Reuter et al. [50], participants (n = 131) were allocated to three different treatment groups and the desired target temperature was about estimated 39–40 °C. In group A1, with 44 participants, the bacterial extracts Serratia marcescens + Streptococcus pyogenes or Pseudomonas aeruginosa were used. In group A2 62 patients were treated with the same bacterial extracts, preceded by 30 min WBH (IRA 1000, Von Ardenne). In group B, containing 25 participants, therapeutic fever was induced by the application of combinations of approved drugs (Colibiogen, Iscador, Picibanil, Polyvaccinum forte, Strovac) and preceded by WBH (n = 25). Even though the authors conclude a reduction of side effects of some bacterial extracts through the preceding treatment with hyperthermia, no statistical data are given for this conclusion.

The drawbacks of this study are low and incomplete reporting regarding basic demographic data and the allocation procedure. Furthermore, the patient collective is very heterogeneous and perhaps also selective. Therefore, we classify the methodological quality of the study as unacceptable and refrain from further discussion of the results in the following.

Endpoints in the second level of evidence (single-arm studies, case series and case reports)

Further information to the individual results of each study is shown in Table 4 (outcomes in single-arm and case series) and Table 5 (outcomes in case reports).

Endpoint tumour response, pain, quality of life and fatigue

Three single-arm studies exist, which used alternative hyperthermia alone. In these trials no clinical tumour responses were documented [67, 72, 78]. In 22 single-arm studies and in one case series, alternative hyperthermia was used in combination with CTx or RTx. The incidence of a (complete response) CR ranged from 0% in 13 studies [60,61,62, 67, 72, 73, 75,76,77,78, 80, 83, 84] to 37.5% in a study with B-cell neoplasm with 8 participants [69]. In summary, due to the heterogeneous tumour entities and to the fact that all single-arm studies, which reported an improvement in tumour response, used alternative hyperthermia combined with CTx or RTx, and no conclusion can be drawn, whether the addition of hyperthermia to CTx or RTx has an effect on tumour response.

Pain as outcome was analysed in five studies including 106 patients. A reduction of pain was only seen in the 13 patients with objective tumour response, after a treatment which combined WBH with other treatments. Looking at ECC-WBH as part of therapy, 18% and 21% of the 36 participants reported a reduction of abdominal cancer pain [75]. Moreover, for EH as part of therapy for 33 patients, a decrease in median VAS score was reported [83]. In 23 patients, significant reduction in worst pain, least pain, average pain and current pain (p < 0.001 for all) was reported after treatment and was maintained during the next three months. While at baseline 74% of the patients were still taking analgesics, three months after treatment the rate dropped to 48% [87]. The attention patients received due to the hyperthermia intervention must be considered with respect to subjective outcomes such as pain. Considering adverse events, it has to be noted that pain was also caused by the EH treatment in one study [81] and 20% of the participants refused further EH sessions because of pain [83]. Overall, heterogeneous results on the endpoint pain were documented.

QoL was addressed in six studies including 117 patients. In one study with WBH changes in patient’s well-being were again documented for those patients whose disease responded to therapy [71]. Another study with 22 patients using WBH reported an improvement in QoL [61]. No significant differences in the QoL were seen from baseline to three months after EH with RTx in one study with 10 patients [83] and in another study [82] with initially 19 patients. Another study using EH with 23 participants [87] showed that except for nausea and vomiting, loss of appetite, diarrhoea and financial problems, the patients’ quality of life improved significantly in all the functional scales within three months. Also these results may be explained in part by the attention received during the application of hyperthermia.

One study with 37 patients treated with WBH and CTx reported the outcome fatigue [71]. In total, 34 patients complained about grade one to two fatigue and one patient suffered from grade three fatigue prior to treatment. After treatment, all 16 patients with objective tumour responses reported, as is to be expected, an improved sense of well-being. On the other hand, fatigue is also ascribable to WBH, as reported in five studies [68, 69, 73, 74, 88], or to ECC-WBH, as documented in two studies [78, 89].

Endpoint survival

Survival data were documented in 15 single-arm studies and in one case series. Median overall survival was analysed in 14 studies, but explicit data for OS being calculated from first diagnoses can be found in 3 studies [60, 61, 81]. OS from first diagnoses ranged from 18.6 months [81] to 19.3 months [60, 61]. The TTP (time to progression) or the PFS (progression-free survival) measured in ten studies ranged from 2.5 months [82] to 6.8 months [60, 61]. The 1-year OS was part of analysis in five studies and ranged from 30% [86] to 68% [60, 61]. Due to lead time bias, the data have to be treated with caution. By reason of the design of single-arm studies, no data are available about the lifetime of a control group, so it is not possible to derive a clear statement in which way hyperthermia may influence survival data.

Adverse events

Detailed information to the different side effects can be found in Tables 3 and 6.

Related to WBH

In the SR by Lassche et al. [45], myelosuppression grade 3 and 4 occurred most frequently in studies using WBH along with CTx. Grade 3 and 4 side effects that were directly attributable to WBH therapy included cardiac arrhythmias, dermal side effects and kidney failure. Four patients died of treatment-related complications [48, 60, 66]. In the prospective cohort study by Gerke et al. [49] 43 patients with advanced sarcoma were divided into three groups (ICE (ifosfamide, carboplatin and etoposide)-CTx in combination with extracorporeal WBH (e-WBH), ICE-CTx with r-WBH by infrared radiation (Aquatherm) or only ICE-CTx). In both hyperthermia groups, the time at a target temperature of 41.8 °C was one hour. On the third day of the cycle, the glomerular filtration rate (GFR) decreased significantly more in the WBH groups than in the group treated with ICE-CTx alone and no difference between e-WBH and r-WBH was found (p = 0.364). Three weeks after the start of the CTx-cycle, GFR and the serum creatinine showed no significant difference between the different treatment modalities. In summary, nephrotoxicity sees to be aggravated by WBH immediately after chemotherapy especially when nephrotoxic agents are used along with WBH.

In the RCT by Robins et al. [47] 16 patients with different types of advanced cancer were treated with WBH (Aquatherm) alone during week 1. Thereafter, they were randomized to receive either Melphalan alone in the second week, Melphalan plus WBH for one hour at a target temperature of 41.8 °C in the fifth week or the reverse sequence. Across all CTx levels for Melphalan + WBH the average mean nadir WBC count was 35% and the mean nadir platelet count was 20% lower compared to Melphalan alone (p = 0.006, p = 0.04), denoting that myelosuppression was more pronounced in cycles with WBH.

Lesions of the skin were a frequent side effect. In one study with 9 participants treated with WBH, 60% showed a transient erythema and in one subject two round thermal lesions (grade 2) appeared [46], respectively; 3 of 44 patients showed pressure scores [48]. In the lower class of evidence skin lesions, most pronounced in grade 1 were also a frequent side effect [59, 60, 62, 64, 66, 71,72,73,74, 88]. The lesions included blisters, erythema, burns and ulcers especially at contact points with the heating pads. Besides waterbed and infrared rays, WBH also uses insulating measures and thermal blankets. These can then cause the skin lesions.

According to the authors, herpes simplex infections were attributed to WBH. In one study 39% were detected with mucosal herpes infections, responsive to acyclovir [48] or such infections (grade 1) occurred in 7 of 16 patients [47]. In the lower class of evidence herpes infections were also often reported [65, 68,69,70, 73, 74] and the incidence ranged from 17% [70] to 73% [74].

In the RCT by Robins et al. [47] a transient increase in liver function tests (grade 2) was seen in 3 patients and additionally low-grade fever (< 24 h post-treatment) occurred to 3 patients. While taking a closer look at serum chemistries in the lower class of evidence, liver enzymes showed a significant elevation 24 h post-WBH [74], but returned to normal range at follow-up [74, 88]. Another study reported a transient elevation in liver enzymes in 25% [67].

A fatigue syndrome grade 3 and 4 was noted in a quarter of the patients in cycles with WBH; compared to cycles without WBH, grade 3 and 4 occurred in 9% [48]. In the lower evidence class, fatigue was mentioned likewise [68, 69, 73, 74, 88] and the incidence ranged from 14% [88] to 100% [74]. Transient cardiac arrhythmias with electrocardiographic signs of myocardial ischaemia (WHO grade 3) concerned to 5 of overall 44 patients in the CT by Hegewish-Becker [48]. In single-arm studies, case series and case reports, cardio-circulatory events were reported as arrhythmic episodes [59, 60, 68, 74, 88], depression of blood pressure [68, 70, 73, 74, 88] or an increase in heart rate [67, 73, 74, 88].

In addition, in studies of the lower class of evidence, following adverse events were mentioned, which were related to WBH, according to the authors. After treatment, some patients suffered from slight headache (between grade one and two) [65, 68,69,70, 72]. Neurological adverse events were peripheral neuropathy of the femoral, peroneal and ulnar nerve [73, 88], psycho-motoric dysfunctions (grade one and two) [59], convulsions, hallucination (grade one and two) and severe rhabdomyolysis [73]. Further side effects, related to WBH, included post-hyperthermia fevers, lasting for a maximum of 36 h after treatment [65, 68, 73, 88], urinary tract infections [65, 68, 69], nausea and vomiting at most grade two [67, 68, 73], diarrhoeas between grade one and two [65, 70, 73, 88] renal failure grade one or two [59] and calf thrombophlebitis [68].

Related to therapeutic fever

In one study by Reuter et al., patients experienced nausea and vomiting, headache, back pain, circulatory reactions and weakness the following days [50].

Related to ECC-WBH

Side effects ascribable to that modality were reported in studies of the lower class of evidence. Probably ascribable to ECC-WBH were 4 deaths (24%) due to intra-abdominal bleeding, 1 death due to lung oedema and to 2 due to hepato-renal syndrome [75]. Proteinuria grade 1 to grade 3 was reported in three patients [89]. Transient liver failure occurred in 1 patient [89, 91] and hepatitis was detected in 3 participants [77]. During ECC-WBH alone elevation of bilirubin, albumin, lipase, AST (aspartate aminotransferase), ALT (alanine transaminase), γGT (gamma-glutamyl transferase) could be found [89] or AST, ALT and bilirubin significantly increased 24 h after treatment in another study [91]. Moreover, perioral herpes infections occurred [76, 77, 89]. Skin lesions included burns [89], pressure scores [76, 77] or a skin imprint [78]. Several studies reported side effects which were most probably due to cardiovascular stress such as mild anasarca [76, 77], episodes of hypotension grade two [89, 91] or the significant increase of the heart rate during heating [76, 77, 89, 91]. Moreover, the administration of catecholamines or crystalloid solutions were necessary [76, 77, 89, 91]. Adaptation to warming during ECC treatment may well be associated with circulatory difficulties and is therefore not suitable for a wide patient population. Further side effects included an infection of the shunt necessitated the removal of graft in 1 patient [75], changes in coagulation parameters [77, 89, 91]. Fatigue [89], nausea [77, 89], vomiting [77] and diarrhoea [76, 77] were documented with grade 1 or 2. Moreover, between slight to severe weakness of muscles [75], reversible paresthesia grade 2 [77] and post-treatment fever grade 1 [89] occurred. In the combination ECC-WBH with CTx 11% [76], 8%, respectively, [77] of the participants developed acute renal failure, requiring haemodialysis and nephrotoxicity was reported in 42% of the participants [77]. Neurological problems included encephalopathy in one patient, who died 3 weeks later, but according to the authors, the death was not attributed to treatment [78].

Related to EH

In the RCT by Minnaar et al. [51], adipose tissue burns occurred in 9.5% and pain in 9% of patients treated with EH. In the cohort study by Kim et al. [57] in the first 60 days after treatment, a significantly higher opioid dose was used in the group treated with hyperthermia (p = 0.022). Comparing the arms in the RCT by Loboda et al. [54] which included treatment with and without EH, there were no differences in haematological and gastrointestinal toxicities or liver and kidney function. In the RCT by Mahdavi et al. [55] EH-related side effects were mild headache with no necessity for any additional medication. In the RCT by Kim et al. [58] fat necrosis and hot spots occurred in 1 patient each. Comparing arm A, which received EH in addition to CTx and RTx, with arm B, the number of side effects did not differ significantly, except for gastrointestinal side effects. These occurred significantly more frequently in arm B (p = 0.01). Adverse events caused by EH in the RCT by Fiorentini et al. [56] included headache, scalp burn and seizures. More than an hour after treatment, seizures occurred in 4 additional patients.

In the lower class of evidence, local pain grade 1 occurred to 20%, grade 2 was documented in 60% [81], and in the study by Yu et al., 20% of the participants refused further EH sessions because of pain [83]. Typical symptoms of increased intracranial pressure, e.g. nausea (grade one to two), confusion (grade one to three), somnolence (grade two to three) and focal neurological symptoms, for example aphasia (grade three to four) or hemiparesis (grade two to three), were documented in 1 study [81]. Furthermore, 87% of the patients suffered from headache up to grade 4 in 1 study [81].

Adverse events in studies with multiple interventions

A brief characterization of these studies can be found in Table 7, and information about the adverse events related to the hyperthermia treatment in Table 8.

Related to WBH

The side effects are in accordance with the prior reported adverse events. The incidence of the skin lesions was up to 100% [99, 137] and the severity was up to grade 3 [99, 135, 137]. Herpes infections were a common side effect [99, 126, 127, 137], and the incidence was also up to 100% [99]. Neurological side effects included encephalopathy in overall 3 patients [112, 137], which reversed spontaneously. Moreover slight CNS (central nerve system) dysfunctions or disorientations [112, 137], neuropathy or transitional psychosis [137] was reported. During treatment tachycardia [137] or tachyarrhythmia [112] occurred. Fatigue as an adverse event was noticed in two studies, and the incidence ranged from 24% [126] to 100% [99]. Further side effects included headache [126], haemodynamic depression [137], nausea and vomiting [126, 127], transient elevation in liver function tests [127] and pain [137].

Related to EH

Skin lesions consisted of local redness, subcutaneous fibrosis of fatty tissue and slight skin burns [102, 124, 125, 130, 141]. A small number of participants suffered from short-term asthenia, headache, abdominal or local pain and chest pressure [124, 125, 130, 141]. In the study by Yu et al. 30% of the patients refused further EH sessions, mainly because of pain. To complete treatment, the administration of opioids was necessary [139]. Further adverse events related to EH included vomiting or tachycardia [124].

Discussion

We categorized the studies into two levels of evidence. The higher evidence level one includes SR, RCTs, CTs and cohort studies, whereas evidence level two reports on single-arm studies, case series and case reports.

Studies evidence level one with methodologically acceptable evaluation

Potential benefits of WBH

The SR by Lassche et al. [45] showed no benefit of the invasive WBH treatment. Due to the absence of two-arm studies in the SR, the effect of WBH as addendum on tumour response rate is only speculative. With the numerous grade 3 to 4 side effects, attributable to WBH, the risk/benefit ratio clearly shifts the side of the risks.

Potential benefits of EH

In the RCT by Loboda et al. [54] the 10-year OS in the arm with EH was significantly higher, although no significant difference was seen in tumour response. It must be critically noted with regard to the 10-year OS that no exact numbers are given, but only illustrations, from which no exact data can be read off. In addition, no information about the method of randomization is given. In the RCT by Minnaar et al. [52] the intervention group showed significant improvements in some sub-items (social functioning, emotional functioning, fatigue and pain) of the EORTC questionnaire compared to control. Furthermore, the group treated additionally with EH achieved significantly better values at six-month local disease-free survival, local disease control and tumour response. The reasons for the missing data of part of the participants are not stated; therefore, selective reporting cannot be excluded. Additionally, with such a high drop-out rate and without any reasons given, the comparability of the groups cannot further be assumed. It is therefore possible that healthier or more motivated patients remained in the study. Those patients then may achieve a better result and do not constitute a representative sample [51]. Although, six months after treatment, the tumour volume in the intervention group was lower than in the control group in the RCT by Mahdavi et al. [55], neither OS after 18 months nor the Karnofsky Performance Status Scale showed any benefit. However, due to missing comparison of the patient and tumour characteristics at baseline, it is unclear whether the groups already differed from the beginning and the intervention group may have had better values all along. Therefore, no valid interpretation of these results is possible. In addition, a rationale for the allocation to the two treatment arms is not specified [55].

Studies evidence level one with methodologically low evaluation

Potential benefit of WBH

WBH did not improve the quality of recovery, as no significant difference was found in the RCT by Sulyok et al. [46] for the quality of recovery after surgery overall or in the dimensions assessed (global QoR-40, p = 0.81). On the other hand, despite the fact that in vitro data demonstrated that hyperthermia distinctively enhances the cytotoxic side effects of oxaliplatin [142], and the incidence of toxicities most likely related to chemotherapy and was hardly different between the chemotherapy cycles treated with or without WBH in the phase-II-CT by Hegewish-Becker et al. [48].

Potential benefit of EH

In the retrospective cohort study by Kim et al. [57] the EAS over the time showed worse values in the EH group and a significant higher opioid dose within the first 60 days. In another retrospective cohort study by Kim et al. [58] treating the intervention group with EH, no significant difference was found in the pathologic outcome. Moreover, no difference was seen in 2-year OS. The only significant difference was reported in the EH group in the two-year locoregional recurrence-free survival. In the retrospective cohort study by Fiorentini et al. [56] the intervention group was treated only with EH, whereas the control arm, received best supportive care together with CTx. The overall response rate and the median OS were significantly better in the hyperthermia group, although the difference in survival was only a few months. With regard to the methodology of the study, the results must be considered with caution. No demographic parameters comparing the intervention and control group at baseline as well as information on prior treatments are not specified. Even if a benefit of hyperthermia treatment is first conveyed in these retrospective studies, this cannot be considered an evidence-based benefit due to the methodological limitations of the studies.

Adverse events in evidence level one and well-reported adverse events in evidence level two

Side effects related to WBH/ECC-WBH

The SR of Lassche et al. [45] points to several grade 3 and 4 toxicities, the invasiveness of the procedure, the elaborate supportive care and the high costs. Moreover, the RCT by Robins et al. [47] showed that myelosuppression was more pronounced in cycles with WBH. One major problem of hyperthermia is organ toxicity as nephrotoxicity. Special care should be taken with patients with restricted kidney function [49]. Other reported adverse events included cardio-circulatory stress during WBH or ECC-WBH. So patients with higher grade of cardiac arrhythmias have to be excluded [3] as arrhythmic episodes occurred regularly [59, 60, 62, 68, 74, 88, 137] and heart rate increased [67, 73, 74, 76, 77, 88, 89, 91] during heating. Therefore, cardiologic examinations must be carried out to prove the patients’ cardio-pulmonary capacity before treatment and continuous cardiac monitoring during WBH is necessary [62]. To maintain a sufficient blood pressure, crystalloid solutions and/or catecholamines were needed during plateau phase [76, 77, 89, 91]. Furthermore, patients with markedly restricted hepatic capacity have to be precluded [3], because during WBH, hepatic dysfunction was mentioned [60, 66, 70, 127] or a transient elevation in liver enzymes occurred [67]. Another severe side effect is the more pronounced myelosuppression in cycles with WBH [47]. In addition, the application of WBH for patients with cerebral or spinal metastases should be critically reconsidered, because there may be the risk of an increase in intracranial or intraspinal pressure [3, 97]. In particular, the use of ECC-WBH should be reconsidered critically, as due to the high invasiveness 4 deaths from overall 76 patients treated were attributed to ECC-WBH [75]. Additionally, elevation of liver enzymes occurred [89, 91] and participants developed acute renal failure, requiring haemodialysis [76, 77]. Moreover, most of the patients needed analgo-sedation or deep anaesthesia during WBH or ECC-WBH [3].

Side effects related to EH

A somewhat milder side effect profile was seen with EH treatment. EH-related side effects were mild headache with no necessity for any additional medication [55]. Other adverse events included fat necrosis or hot spots [51, 58], scalp burn, seizures [56] and pain [51]. In the cohort study by Kim et al. [52] a significantly higher opioid analgesic dose was used in the group treated with hyperthermia. (p = 0.022) or participants refused further EH sessions because of pain [83].

Risk/benefit ratio

When considering the risk/benefit ratio, it becomes apparent that, due to the very heterogeneous results and methodological limitations of the included studies, clinical evidence for the benefit of alternative hyperthermia in cancer patients is still lacking. Based on the current research, the adverse events outweigh the potential but yet unproven benefits of alternative hyperthermia.

Absence of intra-tumoural temperature measurement in alternative hyperthermia

Moreover, the term hyperthermia is misleading because it is not clearly defined. Based on the rationale behind hyperthermia, some clinical studies with conventional hyperthermia were able to show evidence-based benefits for selected types of cancer [143]. These hyperthermia treatments were carried out with defined quality standards, including an intra-tumoural temperature measurement every minute, a exactly determination of the treatment area previous via MR (magnetic resonance) or CT (computer tomography) or the presence of an engineer or physicist during treatment [144]. Invasive temperature probes represent the gold standard in thermometry. For non-invasive monitoring, CT-, MR- and ultrasound-based thermometry methods have been developed although these do not deliver the same accuracy [145, 146]. Exact temperature control is essential in hyperthermia, on the one hand to avoid side effects [147], on the other hand to reach the desired target range, with regard to the close dose–effect relationship [145, 148]. Provider of alternative hyperthermia concepts also advertise with the same mentioned theoretical principles, but the implementation of uniform rules is missing. In the present review, information about temperature measurements is lacking in 15 studies [46, 50, 55, 57, 58, 64, 81, 82, 84, 85, 87, 94,95,96,97]. In the studies by Minnaar et al. [51,52,53], the authors do not consider it necessary to measure the temperature. In only one study with electro hyperthermia thermal mapping was performed in accordance with the guidelines of the European Society of Hyperthermic Oncology, according to the authors, but further information about the exact way of measuring temperature is missing [25,26,27,28,29,30, 80]. In one study WBH temperature was measured constantly by a probe placed in the centre of the tumour [93]. In another study with EH intra-tumoural temperature was approximately determined using a thermal imaging camera [54]. The other studies did not measure the intra-tumoural temperature directly but only offered indirect methods for which clinical data are missing [47,48,49, 59,60,61,62,63, 65,66,67,68,69,70,71,72,73,74,75,76,77,78, 83, 86, 88,89,90,91,92].

Limitations of this work

Some limitations of this systematic review must be mentioned. First, we focused on adults, omitting literature that included more than 20% children as patients. This is however not a limitation as paediatric patients generally form an even more heterogeneous population than adult patients, making solid conclusions about clinical effects of alternative hyperthermia even more challenging. So excluding these categories is not expected to change our conclusions. Also, studies that were not in English or German were also not considered. However, including these would have made our in-depth analysis of content and methodology too challenging.

Conclusion

No clear statement regarding the efficacy of hyperthermia treatment in complementary medicine on cancer patients may be derived from published studies. Further randomized controlled trials are necessary, which compare groups treated with standard cancer therapy to groups treated with alternative hyperthermia along with standard cancer therapy to draw conclusions whether alternative hyperthermia influences tumour response, survival data or the quality of life and to find out, which side effects are exactly assignable to hyperthermia. Due to this and to the heterogeneous results of the systematic review regarding the outcomes pain and quality of life, no benefit of alternative hyperthermia has been shown and no evidence-based indications can be stated. The adverse events especially of WBH and ECC-WBH may overweigh the potential benefit. Physicians should not prescribe WBH or ECC-WBH in case of comorbidities like renal or hepatic diseases, cardiac arrhythmias, cardia-pulmonal insufficiency, increased intracranial or intraspinal pressure or existing aspects impeding the essential analogue sedation or anaesthesia.

Due to the highly different methods offered with the same terminus hyperthermia, it is especially difficult for patients to distinguish between scientifically proven hyperthermia treatments and alternative hyperthermia methods.

To help patients and physicians who are not experts in oncology, institutions which offer or evaluate conventional hyperthermia treatments should clearly differentiate their procedures from the offers of alternative providers in words also comprehensible for patients.