Regarding the status of this text: This critical commentary was written two years after the protocol was drawn up and was not involved in the study’s planning. It describes what, from the perspective of outpatient practice, should have been designed differently. None of the points mentioned here were taken into account in the protocol.
The study in brief
Title: Individualised homeopathy to reduce the use of antibiotics in women with recurrent uncomplicated urinary tract infections. The sponsor was the Klinikum rechts der Isar at the Technical University of Munich; responsibility lay with the Department of Nephrology; statistics were handled by the in-house Institute for Artificial Intelligence and Informatics in Medicine; and the protocol committee was based at the Institute for Integrative Medicine at the University of Witten/Herdecke (EudraCT 2021-002214-14).
Study design: A single-centre, randomised, double-blind, placebo-controlled Phase IV trial involving 240 women aged between 18 and 65 years, with 120 in each group; planned treatment duration of nine months per participant.
Treatment: One of 140 homeopathic remedies (Gudjons Pharmacy) in C200 or C1000 potency, five globules administered sublingually at the start of the study, repeatable up to three times and during acute infection; the control group received non-impregnated sucrose globules.
Primary endpoint: The number of urinary tract infections treated with antibiotics between the start of the study and month nine. An infection was defined as per the German S3 guideline: clinical symptoms plus bacterial count in a urine culture or plus a positive test strip.
Sponsor and funding: The Bavarian State Government initiated and funded the study in collaboration with the Technical University of Munich.
The study was discontinued because too few participants could be recruited. According to Ulf Riker, 2nd Vice-Chair of the German Central Association of Homeopathic Doctors, fewer than 20 per cent of the required 240 female participants were recruited, i.e. fewer than 48 participants (commentary dated 13 August 2026).
Consequently, neither superiority nor inferiority compared with placebo can be demonstrated; the widespread assumption that the study failed to show any efficacy is therefore incorrect.
The real issue at the heart of the criticism: what is actually being treated here?
The crux of the criticism is not the statistics, but the diagnosis. Under the presenting complaint of ‘recurrent cystitis’, four different clinical presentations occur in outpatient practice, which overlap and vary greatly in their individual course. The study summarises them under a single definition and a single criterion for progression. The criterion for comparison is the frequency of antibiotic administration (antibiotic response).
|
Symptoms |
Common characteristics |
Urinalysis findings |
Antibiotic response |
|
Recurrent cystitis |
Sudden, painful urge to urinate, cloudy urine; sleep is regularly disturbed; up to twelve episodes a year are not uncommon |
Leucocytes, often nitrite-positive, cloudy and malodorous |
Yes, improvement within 1–2 days, findings resolved after five days |
|
Irritable bladder |
Persistent urge to urinate without night-time symptoms (“Sleep relieves me of the stress”), burning sensation at the end of urination; triggers include anxiety-inducing situations and avoidance behaviour |
Variable, slightly elevated white blood cell count, nitrite usually negative, sometimes weakly positive |
No, multiple routine treatments without clear effect |
|
Urethral syndrome |
paroxysmal, cramp-like burning pain at the urethral orifice lasting for hours, tendinous insertions tender to pressure, symphysis pubis tender to touch |
usually unremarkable, occasionally weakly positive with no temporal correlation to the pain |
No, but is nevertheless routinely treated with antibiotics |
|
Interstitial cystitis |
Pressure pain when the bladder fills, increasing over the years; extremely frequent urge to urinate; relief after voiding; flare-ups triggered by diet and prolonged physical strain |
Uncertain, borderline, with no clear correlation |
only a short-term or partial response to antibiotics |
The implication for the study: only the first of the four clinical presentations responds to antibiotics in a way that resolves the crisis. If antibiotic use is chosen as the primary endpoint, this measures a treatment for irritable bladder, urethral syndrome and interstitial cystitis whose efficacy is questionable depending on the phase of the condition. If the differences between the four clinical presentations are not clearly documented, it is impossible to determine which participant, with which primary symptoms, ended up in which comparison group.
Objections to the design – eight points left unaddressed in the protocol
1. The time frame contradicts the inclusion criteria. Participants are included if they have had at least two infections in six months or three in twelve months. The study is conducted over nine months. Any difference would have to arise during this period from just a few single doses of a homeopathic remedy. This presupposes that these single doses, as the sole influencing factor, prevent the onset of the condition.
2. The measurement accuracy is insufficient for borderline findings. Test strips vary; however, a calibratable device would be required for interpretable results. Particularly in cases of irritable bladder and urethral syndrome, findings are often borderline, which is why a colour-matching strip is too imprecise in clinical practice.
3. The inclusion and exclusion criteria exclude the general practice population. Excluded are women whose suitable medicine is not on the provided list of medicines; pregnant and breastfeeding women; postmenopausal women who have not tried local oestrogen – in other words, all those who do not want or cannot tolerate hormones; and women who have undergone homeopathic treatment in the preceding six months. It is precisely those with recurring symptoms who are likely to have tried something during this period. No mention is made of women taking tamoxifen or aromatase inhibitors; nor are oral contraceptives mentioned, even though they can improve or worsen the course of the condition.
4. The study’s list of remedies is incomplete. Compared with over 400 of our own prescriptions, around 10 per cent of the commonly indicated homeopathic remedies are missing from the study design. For example: for the clinical symptom ‘burning sensation when urinating’, the repertory lists around 180 remedies; approximately 25 per cent of these are not included in the study list. However, a suitable homeopathic remedy cannot be replaced by another in individual cases.
5. The timing of administration is not defined. In acute cases, the symptom picture can change from day to day; in highly acute cases, it can even change within hours. A remedy prescribed at midday is appropriate at that time; if, for logistical reasons, it is not taken until the next day or the day after, the case has usually progressed further and requires a new analysis.
6. The methodology itself is not described. ‘Individualised homeopathy’ (iHom) does not specify a standardised procedure; there are several approaches, the comparability of which has not been clarified. The protocol should therefore include the methodology of case analysis, the criteria for case management and, above all, verifiable criteria for determining when a remedy reaction is deemed to have occurred and when a contextual factor is deemed to have been ruled out.
7. The psychosomatic perspective is missing. In young women with recurrent cystitis, sexual conflicts are frequently the underlying cause, which only become apparent after months of treatment; in cases of overactive bladder, a compulsive need for control and repressed aggression are often found; in cases of urethral syndrome, life situations involving excessive strain during life crises—such as the exhausting care of relatives or job loss—can be observed. In some cases, several contextual factors coincide. Such stressors have a direct effect on the immune defences in the affected area. Quality-of-life questionnaires capture these factors only indirectly, at best, and do not constitute a diagnostic criterion for the choice of medication.
8. The rationale for the study and its design are inconsistent. The protocol assumes a psycho-neuro-immunological approach to the effect. However, psychoneuroimmunology has shown that a simple before-and-after design is of little value when dealing with complex clinical presentations. This fact is not taken into account in the design.
Why the termination was no coincidence
The study was terminated due to a lack of participants. The exclusion criteria target precisely those women who present with this concern in clinical practice: those who have undergone unsuccessful prior treatment, pregnant and breastfeeding women, postmenopausal women who have not undergone hormone therapy, and all those whose medication is not on the list. Anyone who excludes the clinical population in this way will not find them again during recruitment attempts.
Thus, the termination of the study itself is a finding rather than an operational mishap: overall, there was a mismatch between the question, the procedure and the subject matter, and this was first evident in the number of participants.
What a constructive approach requires instead
- An analysis of real-world treatment data prior to the design phase, to determine which clinical presentations occur and with what frequency, how long the courses of treatment typically last, and which objectives (can or should) be pursued in individual cases.
- The delineation of the four clinical presentations as inclusion criteria and as evaluation characteristics, rather than a single guideline definition as a uniform standard.
- An observation period appropriate to the recurrence rate and the course of the conditions, and endpoints that also capture what occurs without antibiotics.
- A comprehensive list of medicinal products, or an explanation of why a limited selection can nevertheless answer the question.
- Verifiable criteria for the drug response and for the conditions of the individual case , so that the analysis is not based on just a few proxy variables.