Effects of serial thermal brine hand baths on hand osteoarthritis: a randomized, double-blind, sham-controlled trial

Abstract

BACKGROUND

Balneotherapy is used since ancient times, effects on hand osteoarthritis however are not well elucidated.

AIM

To evaluate therapeutic effects of serial thermal brine hand baths (tbhb) in patients experiencing painful hand osteoarthritis (HOA).

DESIGN

Randomized, double-blind, sham-controlled monocentric clinical trial.

SETTING

Outpatient clinic, monocentric.

POPULATION

Patients with painful bilateral painful HOA meeting the 1990 American College of Rheumatology classification criteria with a baseline pain level >30 mm on the Visual Analog Scale (VAS) and radiographic evidence of hand osteoarthritis in posteroanterior X-ray projection, classified as Kellgren and Lawrence grade ≥2.

METHODS

Probands were randomized to the intervention group (IG) receiving tbhb or the control group (CG) receiving tap water hand baths (sham). Both hand baths (tbhb and sham) were performed 6 times in 3 weeks with 15 min/bath at a water temperature of 25 °C.

OUTCOMES

The primary outcome was the change in pain intensity (VAS [mm]) following the intervention. Secondary outcomes were changes in the Australian Canadian Osteoarthritis Hand Index (AUSCAN), grip strength, functional parameters of microcirculation (by O2C method), vascular endothelial growth factor (VEGF) and safety. Outcomes were primarily analysed after the last intervention followed by secondary analysis of the follow-up visits 4 and 12 weeks after intervention and during intervention.

RESULTS

A total of 38 patients successfully completed the trial. Pain intensity significantly decreased due to tbhb (P<0.001), with sustained benefits observed up to 3 months post-intervention. The AUSCAN pain subscore was significantly reduced (P<0.001). There was no significant change in other AUSCAN subscores, nor in grip strength or functional parameters of microcirculation including VEGF due to tbhb. No adverse events due to tbhb were observed.

CONCLUSIONS

Thermal brine hand baths represent a simple, effective, and well-tolerated method to achieve pain relief in HOA with results lasting up to 12 weeks after intervention.

CLINICAL REHABILITATION IMPACT

Thermal brine hand baths can be used to reduce pain in HOA.

Osteoarthritis of the hands (HOA) is a common, painful and disabling musculoskeletal disorder, with prevalence increasing with age.¹ Clinically, it manifests with bony enlargements, deformities of the hand joints, and occasionally soft tissue swelling.² HOA significantly reduces quality of life, causing hand pain, stiffness, functional limitations, and decreased grip strength.³ As no causal pharmacological intervention exists, treatment focuses on symptom management and involves both pharmacological and non-pharmacological approaches.^{1, 4}

Nonsteroidal anti-inflammatory drugs (NSAIDs) are the recommended first-line pharmacological therapy.^{1, 4} However, NSAIDs should only be used for a limited duration with attention to the risk-benefit ratio. NSAIDs are associated with potentially life-threatening side effects, particularly in the gastrointestinal and cardiovascular systems, especially in older patients.^{5, 6} Consequently, NSAID use is often restricted, prompting patients with HOA to seek alternative treatment options.

While non-pharmacological therapies are internationally recommended, evidence supporting some of these interventions remains limited.^{1, 4} In this context, the previously recommended balneotherapy for HOA^{7, 8} was withdrawn in 2018 by the European League Against Rheumatism (EULAR), citing its reliance on expert opinion and extrapolation from studies on hip and knee osteoarthritis.¹

However, balneotherapy with thermal mineral water is used since ancient times to relief symptoms of osteoarthritis (OA).⁹ Nowadays, balneotherapy for rheumatic diseases has a robust scientific foundation, supported by controlled studies.^10-12 Recent findings reveal intriguing molecular effects of balneotherapy on the cytokine milieu and the immune system.^{13, 14} This is of particular interest as HOA displays inflammatory findings.³ In addition, many studies investigating balneotherapy and/or thermotherapy suffer from methodological limitations, highlighting the need for new clinical trials employing more rigorous study designs and statistical analyses.

Therefore, we conducted a randomized, prospective, double-blind, sham-controlled study with a 12-week follow-up to evaluate the effects of serial Bad Nauheim thermal brine hand baths in patients with painful HOA.

Materials and methods

Ethical approval was granted by the ethics committee of the Faculty of Medicine of the Justus-Liebig-University Gießen, Germany (vote no. 56/22). Written informed consent was obtained from all participants. All the examinations comply with the Helsinki Declaration.

We conducted a prospective, randomized, double-blind, sham-controlled, monocentric trial with a parallel design and a 1:1 allocation ratio to assess effects of serial thermal brine hand baths (thermal) vs. tap water hand baths (tap) on HOA. Two care providers were blinded to group assignments and delivered interventions in both groups. No changes to the methods, including eligibility criteria, were made after the trial commenced.

For inclusion patients with bilateral painful HOA had to meet the 1990 American College of Rheumatology (ACR) classification criteria.¹⁵ Eligible patients were required to have a baseline pain level >30 mm on the Visual Analog Scale (VAS) and radiographic evidence of hand osteoarthritis in posteroanterior X-ray projection, classified as Kellgren and Lawrence grade ≥2.¹⁶ Patients were excluded if they were taking analgesic therapy (NSAIDs/COX-2 inhibitors, metamizole, opioids), steroids, glucosamine medication or received physical therapy for HOA 12 weeks prior to study start. Patients had to have no prior therapy with hand-bath (ever). Furthermore, patients were excluded if they had (secondary) HOA due to various inflammatory rheumatic diseases or active HOA (acute inflammation at study start), trauma or open wounds. Patients with diabetes, vasoactive medication, trophic skin disorders, concomitant vascular disease, sensory deficits (polyneuropathy), Raynaud disease or smokers were excluded.

The study was conducted at Campus Kerckhoff, Justus-Liebig-University Gießen, Dept. of Rheumatology, Clinical Immunology, Osteology and Physical Medicine, Bad Nauheim, Germany. All data was collected at Sprudelhof Bad Nauheim, where interventions were delivered.

The intervention group (IG) received serial thermal, the control group (CG) received a sham intervention consisting of serial tap. Both groups received 6 hand baths in 3 weeks with 2 hand baths per week. Each hand bath had a duration of 15 min. The temperature of 25 °C was the same for both thermal brine and tap water to ensure blinding. Temperature was chosen based on the natural temperature of the thermal brine, which was not heated artificially. Thermal used Bad Nauheimer thermal brine (sodium chloride hydrogen carbonate acidulant: CO₂ 2250 mg/L, HCO³ 1068 mg/L, Ca²⁺ 1370 mg/L, Cl^- 18100 mg/L, Na⁺ 10600 mg/L, test report Institut Fresenius No. 6060087, 09/2022, Bad Homburg, Germany). Tap used local tap water (CO2 n.a. mg/L, HCO₃ 176 mg/L, Ca2+ 62.3 mg/L, Cl- 70.8 mg/L, Na+ 32.8 mg/L, Prüfbericht 250000989, Stadtwerke Bad Nauheim GmbH). Interventions were administered after a 20-minute acclimatization period (pre-rest phase) in a quiet and air-conditioned room at Sprudelhof Bad Nauheim (23 °C). Both thermal and tap were prepared according to protocol in identical bowls in a designated preparation room by unblinded study staff before being brought to the study room. The hand baths were covered, ensuring that both care providers and the patients remained blinded to the treatment allocation. Before immersion an optical sensor was attached to the end phalanx of the index finger of the utility hand for assessing microcirculation (by O2C). Since effects of thermal baths on skin circulation may vary at different times of the day, the immersions were always carried out in the period from 9 to 12 a.m. (Monday to Friday).¹⁷ Interventions were administered by protocol to enhance adherence of care providers and participants. Due to the COVID pandemic and regulatory requirements only one patient at a time received the intervention. Patients did not meet.

Primary outcome of this trial was pain relief after the intervention period. Pain relief was defined as the change from baseline in resting pain intensity after intervention. Pain (primarily at rest and secondary during activities of daily life within the last 48 hours) was assessed on a standardized VAS (0 mm=no pain, 100 mm=worst possible pain) before and after each intervention. In addition, follow-up assessments evaluated pain 4 weeks (w4) and 12 weeks (w12) after the intervention period. A change in pain greater than 10 mm on VAS in general¹⁸ and 20 mm on VAS measuring chronic musculoskeletal pain¹⁹ intensity is defined as the minimal clinically important difference (MCID).

Secondary outcomes were changes from baseline 1) in the Australian/Canadian Osteoarthritis Hand Index (AUSCAN), 2) grip strength, 3) microcirculation and 4) patient satisfaction as well as 5) number and manner of adverse events. All secondary outcomes were assessed at every intervention and at both follow-up visits (w4 and w12).

The AUSCAN is a validated, self-reported outcome measure specifically designed to assess symptoms and functional limitations HOA. It comprises three subscales: pain (5 items), stiffness (1 item), and physical function (9 items). Participants rate each item on a Likert scale (e.g., 0=no symptoms, 4=extreme symptoms), with higher scores indicating greater severity. The total score is the sum of the subscales, providing a comprehensive evaluation of the impact of HOA on daily life.²⁰

Grip strength was assessed using an electronic hand dynamometer (Model No.: EH101, manufacturer: Kuptone, MAX 90 kg), a device comparable to the Jamar Dynamometer, with high reliability and validity demonstrated in studies (interrater reliability: r=0.97; test-retest reliability: r=0.9).^{21, 22} The device measures isometric hand muscle strength in kilograms. Participants were instructed to grip the device with maximum effort. Initially, a “warm-up phase” consisting of three measurements per each hand was conducted, followed by three additional measurements, with the mean value calculated to enhance re-test reliability. This method provides a quantitative evaluation of hand function and strength, commonly used in HOA research.^{23, 24}

Microcirculation and tissue oxygenation were assessed using the O2C (Oxygen to See) device (LEA Medical Technology, Gießen, Germany), a validated method combining photospectrometry and laser Doppler flowmetry.^{25, 26} An optical sensor placed on the index finger of the dominant hand measured three parameters without artifacts: microvascular blood flow (arbitrary units, AU), postcapillary oxygen saturation (SO₂, %), and relative hemoglobin concentration (rHb, AU), representing vascular density and microvascular filling. This approach provides a comprehensive evaluation of functional microcirculatory parameters.^{21, 22, 25-27}

In addition, Vascular Endothelial Growth Factor (VEGF), a key regulator of angiogenesis and new blood vessel formation²⁸ was measured in serum samples collected before the first and after the last (6th) hand immersion ensuring consistent timing to control for circadian variation. Serum was quickly delivered to the laboratory to be centrifuged at 3,500 rpm at 15 °C for 10 minutes, stored at -80 °C, and analyzed using Quantikine^® ELISAs (R&D Systems, Minneapolis, MN, USA; range: 31.3-2000 pg/mL) following the manufacturer’s instructions.

Patient satisfaction regarding the effectiveness and importance of the treatment was assessed after the final (6th) hand immersion using a Likert Scale, a validated psychometric tool.^{29, 30} Participants rated their experience on predefined statements: effectiveness (“not effective” to “very effective”) and importance (“not important” to “very important”), providing a structured evaluation of their perceived treatment outcomes.

There was no change in trial outcomes after the trial commenced.

The sample size was determined to be 25 patients per group with a drop-out rate of 20% to demonstrate the superiority (difference of 15 mm VAS) of thermal compared to tap based on the primary outcome (change in pain levels). Effect sizes were extrapolated as no comparable trials are published.³¹ Calculations were performed using nQuery 8.4.1.

Randomization was applied by a random allocation sequence that was locally generated using a computerized algorithm. A blocked randomization approach with a block size of 2 was used to maintain balanced group sizes. The random allocation sequence was generated by a designated statistician who was not involved in patient enrollment or treatment administration.

This study was conducted as a double-blind trial by protocol ensuring that participants, study personnel, and outcome assessors were blinded to the intervention allocation. No co-interventions were administered during the study.

Population descriptives are presented as means with standard deviations (SD) for continuous variables and percentages for categorical variables. Treatment effects are reported with standard errors (SE) and 95% simultaneous confidence intervals (CI), as appropriate. Differences in outcomes between groups at discharge and follow-up were assessed using linear regression, with baseline values included as covariates to account for initial differences. An Analysis of Covariance (ANCOVA) was employed to compare group means while statistically controlling for baseline values. All analyses were done using a per-protocol approach. The statistical analysis followed a hierarchical approach to reduce multiple testing. The primary focus was to analyze the difference between the IG and the CG after completing the intervention (6 hand baths over 3 weeks). This analysis addressed the main hypothesis of the study and was the highest priority. If a significant difference was observed in the primary analysis (confirmatory), further analyses were conducted at the follow-up visits (4 weeks and 12 weeks post-intervention) (confirmatory). Should these analyses also show significant differences, exploratory analyses were performed to assess differences between groups after each individual intervention (hand baths 1, 2, 3, 4, and 5).

No clustering by care providers or centers occurred in this study; thus, no adjustments for clustering were required. No subgroup analyses or additional adjusted analyses were conducted. Calculations were performed using R version 4.4.1 for Windows, alongside RTools 4.4 and RStudio (Version 2024.09.0+375; Released: 2024-09-23).

Data availability statement

The data associated with the paper are not publicly available but are available from the corresponding author on reasonable request.

Results

Sixty-two patients were assessed for eligibility, fifty were enrolled, randomized and equally allocated to either the IG or the CG (sham). The IG received serial Bad Nauheim thermal brine hand baths (thermal). The CG received a sham intervention consisting of serial tap water hand baths (tap). Thirty-eight patients completed the whole trial and were analyzed. Only participants who completed the intervention and provided outcome data (N.=38) were included in the analysis, as no post-baseline data were available for early discontinuations (N.=12). The study was conducted over a 6-month period, from September to February 2023, amidst the ongoing COVID-19 pandemic (Figure 1).

Figure 1.

—CONSORT flow chart.

Baseline patient and disease characteristics are shown in Table I, II. Mean age was 72 years with most participants being female (N.=43, 86%). Mean disease duration (time since diagnosis) was 8 years with a symptom duration (mean) of 14 years (Table I, II).

Table I. —Baseline characteristics.

	Amount (N.) and percentage (%) or mean±SD
Sex
Male	7 (14%)
Female	43 (86%)
Age (yrs)	72±9
Symptom duration (yrs)	14±5
Time since diagnosis (yrs)	8±3
Clinical HOA
Heberden nodules (N. of patients)	36 (72%)
Bouchard nodules (N. of patients)	29 (58%)
Occupational status
Employed	16 (32%)
Retired	34 (68%)
Radiographic hand OA KL ≥ 2 (N.)	50 (100%)
Generalized OA (N.)
Knee OA	18 (36%)
Hip OA	10 (20%)
Knee and Hip OA	12 (24%)

OA: osteoarthritis; HOA: hand osteoarthritis; KL: Kellgren Lawrence.

Table II. —Clinical baseline characteristics.

	IG	CG
Male/female (N.)	4/21	3/22
Self-reported health-related quality of life (mean±SD)
AUSCAN pain	11.8±2.6	11.3±3.0
AUSCAN stiffness	2.3±1.3	2.16±1.4
AUSCAN function	20.5±9.62	21.9±10.3
AUSCAN total score	31.7±14.9	33.5±15.8
Self-reported hand pain (N.; %)
None	0	0
Mild	6 (32%)	6 (32%)
Moderate	11 (58%)	11 (58%)
Severe	2 (10%)	2 (10%)

AUSCAN: Australian/Canadian Osteoarthritis Hand Index; IG: intervention group; CG: control(sham) group.

The primary endpoint of the trial was met. After the intervention period the CG reported significantly higher VAS values (mean: 33.4, SE: 1.81, 95% CI: 29.8-37.1) compared to the IG (mean: 7.08, SE: 1.81, 95% CI: 3.4-10.8). The difference between groups was 26.4 points, which was statistically significant (SE: 1.81, 95% CI: 29.8-37.1; P<0.0001). The difference of 26.4 points on the VAS exceeds the minimum clinically important difference (MCID) of 20 mm.¹⁹

Auxiliary analysis showed significant differences between the IG and CG after the first hand bath. Pain reduction increased with each application, especially in the IG. At both follow-up visits (4 weeks and 12 weeks after intervention) the IG had significantly lower pain levels in comparison to the CG (Table III).

Table III. —Change in pain intensity measured by VAS.

		IG (N.=19)		CG (N.=19)		Group difference	P value
		Mean	SD	Mean	SD
Intervention period	baseline	39.47	9.99	39.21	8.47	0.26
	After 1. intervention	33.16	8.62	38.42	8.89	-5.26	0.0017**
	Before 2. intervention	30.26	8.96	39.47	9.99	-9.21	<0.0001**
	After 2. intervention	28.95	8.52	38.95	9.68	-10	<0.0001**
	Before 3. intervention	25	9.03	38.95	11.65	-13.95	<0.0001**
	After 3. intervention	22.37	6.36	38.42	12.25	-16.05	<0.0001**
	Before 4. intervention	22.63	9.09	36.95	10.3	-14.32	<0.0001**
	After 4. intervention	18.95	9.54	34.47	11.11	-15.53	<0.0001**
	Before 5. intervention	15.37	8.63	35.11	7.86	-19.74	<0.0001**
	After 5. intervention	10.11	7.99	35.26	9.24	-25.16	<0.0001**
	Before 6. intervention	9.74	5.73	34.21	9.36	-24.47	<0.0001**
	After 6. intervention	7.11	5.53	33.42	9.2	-26.4	<0.0001*
Follow-up	4 weeks after intervention	8.44	9.31	34.74	19.3	-26.3	<0.0001*
	12 weeks after intervention	9.38	7.26	36.32	19.66	-26.94	<0.0001*

*P values confirmatory; **P values exploratory.

The AUSCAN pain subscore showed a significant reduction in pain (mean difference: 5.86, SE: 0.976, 95% CI: 3.88-7.84, P=<0.0001) in the IG (mean: 5.99, SE: 0.698, 95% CI: 4.57-7.41) vs. the CG (mean: 11.85, SE: 0.679, 95% CI: 10.47-13.23) after the interventions period, which supports the result of the primary outcome parameter. After 4 weeks and 12 weeks of follow-up and no intervention, there was no statistically significant difference between the two groups measured by AUSCAN. During the intervention period no significant differences in other AUSCAN subscales (function, stiffness, total score) were seen (Supplementary Digital Material 1: Supplementary Table I).

Regarding grip strength, measured using a hand dynamometer, both groups demonstrated a trend towards improvement for both hands. Statistically significant differences between groups were not observed, neither during the intervention period nor the follow-up (Supplementary Digital Material 2: Supplementary Table II).

Parameters of microcirculation and tissue oxygenation (microvascular blood flow, postcapillary oxygen saturation [SO₂], and local hemoglobin content) or VEGF levels showed no statistically significant differences between groups during the intervention nor the follow-up period (Supplementary Digital Material 3: Supplementary Table III).

The IG showed more favorable outcomes compared to the CG across all domains. In the general assessment, 13/18 (72%) of participants in the IG rated the intervention as “Very Good” or “Good,” compared to 10/19 (53%) in the CG, with one participant in the CG (5%) rating it as “Insufficient.” For pain relief, 13/18 (72%) in the IG rated it as “Very Good” or “Good,” while this was reported by only 6/19 (32%) in the CG (Supplementary Digital Material 4: Supplementary Table IV).

No adverse events (AEs) or serious adverse events (SAEs) were reported by trial participants. The trial took place during the COVID-pandemic. There were 4 acute SARS-CoV2-infections in the IG and 5 in the CG. All infections were not deemed to be linked to the intervention.

Discussion

To the best of our knowledge, this is the first sham-controlled trial of thermal brine or thermal water hand baths in HOA. While balneotherapy is established in treating knee OA,^{29, 30, 32-37} it’s role in treating HOA remains unclear.

Patients treated with tbhb experienced a significant reduction in pain compared to patients treated with sham (tap water hand baths of the same temperature). Pain reduction was already significant and clinically relevant after the first intervention. Pain reduction was greater with every intervention and thus the primary endpoint (after 6 hand baths) was reached, showing pronounced statistical significance but also clinical relevance with a mean difference of 26 mm on VAS compared to sham.¹⁹ Follow-Up visits 4 weeks and 12 weeks after the intervention without any further physical or pharmacological treatment showed lasting effects with significantly lower pain levels in favour of the IG. Not only pain levels measured on VAS but also AUSCAN pain subscores were significantly lower after the intervention period in patients receiving tbhb. Thus tbhb seem to be a valuable treatment option in rehabilitation for HOA. Due to the rapid pain reduction tbhb seem a valid alternative to NSAID therapy with it’s possible life-threatening side effects.^{5, 6}

When regarding subjective and individual perspectives, patients in the tbhb assessed interventions more favourable than patients receiving sham interventions. Tbhb was evaluated as effective and important.

No AEs were reported, showing tbhb to be harmless.

To investigate possible underlying mechanisms of the clinical effects of tbhb the study incorporated endpoints assessing parameters of microcirculation and tissue oxygenation (microvascular blood flow, postcapillary oxygen saturation and local hemoglobin content measured by O2C) and serum VEGF levels. The mechanisms underlying the pain-relieving effects of thermal brine and mineral water therapy are not yet fully understood. Mechanical, thermal and chemical factors play a role. Heat and water buoyancy help to reduce muscle tension and may increase the pain threshold of nociceptors. Likewise, changes at the molecular level in anti-inflammatory/inflammatory cytokines have recently become known that could explain the pain reduction.^{13, 14} Ingredients from thermal brine water can enter the systemic circulation via the skin, but the pathomechanisms are still largely unknown. Immersion during thermal mineral water therapy has many physiological effects and causes changes in the neuro-immuno-endocrine system.^{38, 39} When looking closely at the differences between tbhb and the chosen sham (covered tap water hand baths of the same temperature) mostly chemical differences appear. Bad Nauheimer thermal brine with its inherent temperature of 25 °C and the decision against artificial heating was used, which contains CO₂. CO₂ can lead to vasodilation allowing better tissue microcirculation and possibly better oxygenation and ultimately less pain with VEGF playing an important role in this interplay.^{40, 41} However, O2C measurements did not show significant differences. In addition, VEGF serum levels were measured as VEGF is discussed as a modulator of pain in OA.^{40, 41} However, VEGF levels did not alter significantly between groups. Albeit, to exclude VEGF or microcirculation as underlying mechanisms of the profound clinical effects on pain of tbhb would be wrong. The time of day may have influenced the vascular response. Although all baths were administered in the morning hours from 9 to 12 a.m., it has been reported that skin perfusion after CO₂ baths peaks in early morning (around 7:00 a.m.) and declines toward noon, with only minor effects observed later in the day.¹⁷ In addition, the circulation-boosting effects of CO2 baths are temperature-dependent and do not occur at bath temperatures below 22 °C and are greatest at around 32-35 °C.⁴² Given the temperature of 25 °C of the used tbhb, the observed effect size may have been attenuated, potentially contributing to the null findings regarding O2C and VEGF, especially in light of the limited sample size. Stronger effects of CO₂ would be expected at higher temperatures or a larger sample size would be needed to observe significant effects on VEGF or O2C when using 25 °C thbh. Both VEGF and O2C were secondary (exploratory) outcomes, the study was primarily powered to detect a difference in pain (VAS).

Non-pharmacological therapy trials impose a special challenge to trail design due to the nature of the physical intervention, which is challenging to blind correctly.^{13, 43, 44} In this trial special care was taken to plan and carry out a sham-controlled trial to ensure the validity of trial findings.^{43, 44} While tap water has been used as a sham for balneotherapy before,⁴⁵ one needs to admit that it may induce more than a placebo. A sham intervention is a control procedure that mimics the physical aspects of the active treatment (such as temperature, immersion, or application method) but omits the presumed therapeutic component (e.g., mineral content, CO₂ or active mud ingredients). This may explain observed effects in the control group. However, when assessing effects and the between-group difference, verum and sham are clearly distinguishable. Thus, the choice of sham seems reasonable.

The study was conducted during the COVID pandemic, which influenced recruitment negatively and may have introduced biases due to restricted access to care. Health regulations during the pandemic limited trial procedures, e.g. only one proband received an intervention at once. Still, acute SARS-CoV2 infection was the major reason for drop-out. In this regard, drop-out in general was more than expected. However, the actual effect size on pain (VAS) was higher than calculated. Therefore, the statistic was not compromised and significant differences could be measured with less probands to analyze. As interventions were carried out only on single probands, unblinding due to communication between probands during interventions was impossible. The study assesses short-term effects over a 3-week application (Supplementary Digital Material 5: Supplementary Figure 1). It remains unclear if effects expand further with prolonged application. However, even 6 applications of TBHH achieved a lasting relief in pain over 12 weeks. Although validated instruments were used to analyze pain, the perception of pain is a complex and highly subjective variable that can be influenced by several factors. In further studies the number and localization of the symptomatic joints should be assessed more closely. Furthermore, future trials should aim to include larger, more diverse cohorts and assess prolonged intervention schedules in a multicenter setting.

The findings of this trial align with broader research on thermal mineral water therapy. Prior studies, particularly in knee OA, have shown comparable benefits for pain relief, function, and quality of life.^{9, 29, 30, 32-34, 46, 47} When considering HOA in particular, two studies have applied somewhat comparable designs and methods . Kovács et al. showed that sulphurous thermal water hand baths (20 min/bath, 15 times over a period of 3 weeks) vs. warm tap water hand baths lead to less pain in the hand, more grip strength of both hands and better Health Assessment Questionnaire Disability Index (HAQ) and AUSCAN scores in patients with HOA.⁴⁸ Horváth et al. showed, that patients treated with either thermal hand baths at 36 °C or 38 °C plus magnetotherapy improved the pinch strength of the right hand and HAQ better than magnetotherapy alone with significant effects over 13 weeks of follow-up.⁴⁵ In comparison to Horváth et al. our study used Bad Nauheimer thermal brine with an inherent temperature of 25 °C. We chose to not heat it artificially. Clinical effects are therefore possibly limited. Indeed, optimal temperatures for CO₂ baths are typically in the indifferent range (32-35 °C).⁴² According to Gutenbrunner et al., microcirculatory responses at 22 °C reach only about one-tenth of those at 32 °C.⁴² In comparison however, we did not use any other intervention that could have influenced or even carried observed effects like Horváth et al. who added magnetotherapy.⁴⁵ When comparing to Kovács et al. one must admit that different interventions were studied albeit both qualify as balneotherapy.⁴⁸

In summary, this trial shows for the first time that in painful HOA serial thermal brine hand baths lead to a rapid pain relief with clinical benefit after the first intervention. After 6 interventions prolonged effects on pain can be observed over a 12 week follow-up. TBBH were without side effects and were rated with a convincing subjective satisfaction in the IG.

Conclusions

This randomized, double-blind, sham-controlled trial demonstrates for the first time that serial thermal brine hand baths provide rapid, clinically relevant, and sustained pain relief in patients with hand osteoarthritis. Pain reduction was evident after the first application and increased with each subsequent intervention, with effects persisting for up to 12 weeks post-treatment. While no significant effects were observed on grip strength, microcirculation, or VEGF levels, the intervention was safe, well-tolerated, and rated favorably by participants. The lack of changes in mechanistic endpoints may be attributed to the moderate bath temperature and limited sample size. Given its simplicity, absence of adverse events, and high patient acceptance, thermal brine therapy represents a promising non-pharmacological adjunct for managing hand osteoarthritis. Future studies should confirm these findings in larger cohorts, explore optimal temperature settings, and further elucidate the underlying mechanisms of action.

Supplementary Digital Material 1

Supplementary Table I

AUSCAN scores

Supplementary Digital Material 2

Supplementary Table II

Grip strength.

Supplementary Digital Material 3

Supplementary Table III

Microcirculation assessment by O2C and VEGF.

Supplementary Digital Material 4

Supplementary Table IV

Likert Scale.

Supplementary Digital Material 5

Supplementary Figure 1

Proband with handbath.