- Email: [email protected]
Nuclear magnetic resonance (NMR) studies of homeopathic solutions
British Homeopathic Journal (2001) 90, 14±20 ß 2001 Nature Publishing Group All rights reserved 0007±0785/01 $15.00 www.nature.com/bhj
ORIGINAL PAPER
Nuclear magnetic resonance (NMR) studies of homeopathic solutions S Aabel1*, S Fossheim2 and F Rise3 Institute of General Practice and Community Medicine, Department of General Practice, University of Oslo, Oslo, Norway; 2Department of Pharmacy, University of Oslo, Oslo, Norway; and 3Department of Chemistry, University of Oslo, Oslo, Norway 1
The ef®cacy of homeopathy is controversial. Nuclear magnetic resonance (NMR) has been used to study homeopathic solutions, showing provocative results. We examined the reproducibility of one of the allegedly positive studies. 1 H NMR spectra were recorded for Sulphur D4, diluted and succussed up to D30 (called potentization) at two different frequencies (300 and 500 MHz). The Sulphur solution had been potentiated according to homeopathic principles with deionized water and alcohol. Water proton T1 relaxation measurements were performed also at 20 MHz for the different potentiated Sulphur solutions. Furthermore, the homeopathic remedy Betula alba 30c (birch pollen extract) and appropriate control solution (deionized water, unsuccussed solutions and placebo globules) were measured analogously, both with frequencies giving spectra and T1 relaxometry. The Sulphur remedies showed identical one dimensional proton spectra (1H NMR) at 300 and 500 MHz, regardless of dilution=succussion stage, from D4 to D30. Furthermore, Betula 30c as a potentiated solution and its controls (ethanol dilutions and Betula diluted but not succussed) showed identical spectra. Nor were there any statistically signi®cant differences in longitudinal (T1) relaxation times between deionized water and Sulphur D10 to D30 preparations. The shorter T1 of the Sulphur D4 preparation could be ascribed to the higher microviscosity within the sample matrix caused by the high concentration of dissolved material. Also, the T1 values of the Betula alba 30c preparation (in globular form) and control placebo globules were identical. In conclusion, published results from NMR research on homeopathy indicating differences between homeopathic solutions and control samples could not be reproduced. British Homeopathic Journal (2001) 90, 14±20. Keywords: homeopathy; nuclear magnetic resonance; Betula 30c; Sulphur D4 ± D30
Introduction The ef®cacy of homeopathy is controversial1 This may be due to the fact that several studies, both clinical and laboratory, apparently have shown positive results without providing plausible mechanistic explanations.2 ± 4 The most controversial point is the very high dilution of the homeopathic remedies. The solutions are in fact produced by a combination of *Correspondence: S Aabel, Institute of General Practice and Community Medicine, Department of General Practice, University of Oslo, PO Box 1130 Blindern, N-0317 Oslo, Norway. Received 10 April 2000; revised 27 June 2000; accepted 22 September 2000
dilution and succussion (called potentization), according to certain speci®c procedures.5 Accordingly, the remedies prescribed by a homeopath may well be diluted so much that there are no molecules of the remedy left in the solution. In theory, the probability of detecting even one molecule of the selected agent is zero or close to zero in homeopathic solutions when the solution is diluted more than 1024 times (ie D24). Several nuclear magnetic resonance (NMR) spectroscopic methods are well suited to study certain kinetic and structural properties of water, both in vitro and in tissue.6 ± 9 Therefore, NMR has been used to examine various homeopathic solutions.10 ± 13 Some results are certainly provocative, reporting a structural difference between the one dimensional 1H NMR
NMR spectra of Sulphur and Betula S Aabel et al
spectra of homeopathic solutions and the unsuccussed solvent of the remedy, ie 87% (v=v) ethanol.12 Since homeopathy is quite popular in the western world,14 and NMR investigations of homeopathic solutions are much easier to perform and assess than clinical studies, we decided to examine the reproducibility of one of the allegedly positive studies.12 Moreover, since a clinical trial on isopathic therapy has earlier been conducted, where the remedy was either a high potency of a birch pollen preparation (Betula 30c) or a placebo preparation,15 samples of this remedy were also investigated by NMR.
Materials and methods Chemicals
Sulphur D4 (diluted 1=104) (dilutions prepared by DCG, Drogesentralen, Gothenburg. Sulphur Merck No 4262458 DAB 7 1992), de-ionized water, 96% (v=v) and 87% (v=v) ethanol, diluted from 96% (v=v) ethanol (Arcus Products, Oslo) with deionized water were used. D2O (deuterium oxide) 99.96% D, was purchased from Cambridge Isotope Labs (CIL Andover, MA). Betula alba 30c tablets (sucrose globules impregnated with a Betula alba 30c solution), placebo tablets (ie sucrose globules), Betula alba 30c dilution (succussed and diluted) and Betula alba (diluted but not succussed) were employed. All Betula alba preparations were made from birch pollen, as described in HomoÈopatische Artzneibuch5 and delivered from DCG. Preparation of homeopathic dilutions
Sulphur. The homeopathic medicine delivered from the producer consisted of pulverized Sulphur with added alcohol and water in a dilution of 1=104 (D4). This solution was diluted with 87% ethanol and further diluted and succussed down to 1=1030 (D30) according to standard homeopathic procedures.5 Brie¯y, 26 unused, capped (plastic) glass ¯asks were washed once with deionized water and ¯ushed with 96% ethanol, dried at 160 C for 1 h, cooled to room temperature and then marked with dilution numbers. Of the initial Sulphur solution, (1=104) 5 ml was pipetted to a glass ¯ask to which was also added 45 ml of solvent (87% ethanol). The ¯ask was closed and vigorously succussed 10 times against an unyielding surface according to homeopathic principles. Then, the cleaned pipette (50% v=v 96% ethanol in deionized water) was reused to transfer an unspeci®ed amount of the solution to a new NMR tube (vide supra) for cleaning purposes (twice), before 2 ml (for relaxometry at 20 MHz), 0.7 ml or 0.4 ml (1H NMR spectroscopy at 300 or 500 MHz, respectively) were transferred from the ¯ask in to the NMR tube. The NMR tubes were immediately sealed with standard plastic caps to avoid further exposure to the surrounding atmosphere. Analogously, the dilution process was
continued, so that a 5 ml sample from the 1=105 solution was transferred to the next container. To this solution, 45 ml of solvent was also added and the mixture was again succussed 10 times. This procedure was repeated until a solution diluted to 1=1030 had been prepared. The dilution=succussion procedures were performed in the air-conditioned NMR laboratories to ensure constant temperature conditions. Every new dilution was assessed with one dimensional 1H NMR measurements to see if a change in the appearance (line width and intensity) of the ethanol signals could be detected. Also, the longitudinal (T1) relaxation times of these potentiated solutions were measured at 20 MHz and 37 C.
15
Betula alba 30c, Betula diluted, placebo and ethanol. The Betula alba 30c remedy, prepared by DCG (as used in experiments reported elsewhere5) was both in liquid and globule from. Betula alba 30c is made from birch pollen. The liquid preparation was succussed and diluted 30 times the solvent being 8% (v=v) ethanol in distilled water. The Betula alba 30c solution was examined at 500 MHz (1H NMR spectroscopy). Betula diluted, a liquid preparation that had only been diluted (with no succussion) 30 times was also examined at 500 MHz. For the T1 relaxation experiments globules of Betula alba 30c were used, meaning that the liquid preparation had impregnated sucrose globules. These Betula alba 30c globules were dissolved in deionzied water prior to the measurements. Placebo globules prepared of pure sucrose were dissolved in deionzied water and measured just like the Betula alba 30c preparation. NMR instruments and ancillary equipment
The one dimensional 1H NMR experiments were performed with Varian XL300 instrument (Palo Alto, CA) (manual) (tuned to 299.944 MHz) equipped with a standard 5 mm 1H=broad band switchable probe and with a Bruker Avance DRX 500 instrument (FaÈllanden, Switzerland) (tuned to 500.13 MHz) that had an inverse triple resonance probe (TXI). The longitudinal (T1) relaxation measurements were performed on a 20 MHz instrument (Minispec PC 120-b Rheinstetten, Germany). The NMR sample tubes used at 300 and 500 MHz were Wilmad Economy (WG5-ECONOMY 5 mm o.d.) tubes (Wilmad Glass Co, Buena, NJ). The samples were investigated with GRtubes (dimension 18010 mm) on the Minispec instrument. NMR methods, computers and software
One dimensional proton spectra (1H NMR) were obtained with the standard pulse program STD1H, delivered with the Varian XL300 spectrometer and acquired and processed with the Varian Magical 6.2B British Homeopathic Journal
NMR spectra of Sulphur and Betula S Aabel et al
16
program in the program in the proprietary Varian computer, equipped with a Motorola 68000 processor. An acquisition time of 3.75 s, pulse width 1 ms (PW 90 25.2 ms), 16 scans (NS) per sample and a delay of 0 s between the scans and the receiver gain, reduced to minimum (manually), were used with the XL 300 instrument. Furthermore, a line broadening (LB) factor of 0.3 Hz was used in the processing. The acquisitions were performed at 20 C on non spinning samples. One dimensional proton spectra (1H NMR) were also acquired with a Bruker Avance DRX 500 instrument and processed on Silicon Graphics Indy computers with the program Xwinnmr (versions 1.3 and 2.1) employing the zg30 (30 degree pulse angle) pulse program. A delay (D1) of one second between each pulse was employed, a total of 16 scans were used for each acquisition, with an acquisition time of 3.17 s used for each scan. The acquisitions were performed on non spinning samples at 25 C, without sweeping the lock channel and without deuterium lock, since no D2O (heavy water) was introduced into the homeopathic samples. A minimum receiver gain of one was used due to the intense 1H signals. A line broadening factor of 1 Hz were employed when processing the 1H one dimensional spectra. The magnetic ®eld homogeneity of the 300 MHz and 500 MHz instruments were optimized (shimmed) on the lock signal with a sample where one part of heavy water (D2O) was added to two parts of 87% ethanol. The same shim parameters were used for each session of successive dilutions. The lock functionalities were then disabled, and the receiver gain on the 1H channels were reduced to a minimum before the actual measurements were performed. No corrections due to variations of the geometric or other properties of the individual NMR tubes were performed between each dilution and each replicate.
1
H NMR spectra of ethanol with different Sulphur dilutions at 300 MHz and 500 MHz
Each potentiated dilution was measured at 300 MHz and every ®fth potentiated dilution was monitored at 500 MHz with one of the NMR instruments as described above. Each dilution investigated was assessed with respect to the intensities and line widths of the 1H NMR signals of ethanol.
1
H NMR spectra of ethanol=water in the presence of different Betula dilutions
The experiments were performed as the Sulphur experiments described above, except that spectra were only recorded at 500 MHz.
British Homeopathic Journal
Longitudinal (T1) relaxation time constants for Sulphur dilutions, Betula 30c globules and placebo globules at 20 MHz
The relaxometric study was performed at 20 MHz to avoid the problems encountered with radiation damping, using high quality probes at higher ®elds (300 MHz and 500 MHz).17 The water proton T1 relaxation times of the Betula alba 30c and placebo preparations were recorded at both 24 and 37 C, with a relaxation delay of 15 and 20 s, respectively. The T1 values of the different Sulphur potentiated dilutions and deionized water were only recorded at 24 C. For each sample, the T1 relaxation time was measured ®ve times. Prior to any measurement, the spectrometer was calibrated by adjusting the ®eld homogeneity, the 90 and 180 pulse widths. The T1 water proton relaxation times were determined from 15 data points, using an inversion recovery pulse sequence.16 As a control of the spectrometer's reproducibility, the T1 relaxation time of an NMR standard (1 mM solution of gadolinium diethylenetriamine penta-acetic acid, diluted Magnevist1, Schering AG, Berlin, Germany) was measured at 37 C. The precision of the instrument was set at 3%. The T1 relaxation data are given as means with their 95% con®dence intervals. One-way analysis of variance was used to test for statistical differences. A probability (P) value less than 0.05 was considered statistically signi®cant.
Results All diluted and succussed solutions of the Sulphur showed very similar one dimensional proton spectra (1H NMR) at 300 MHz, regardless of dilution=succussion stage as did the control dilution without the homeopathic ingredient. In other words, the CH2 and CH3 water (H2O) and ethanol hydroxy (OH) resonances did not change appearance in a physiochemically signi®cant way between each dilution=succussion stage (Figure 1). The minimal differences which we found, were far less than the changes reported by WeingaÈrtner.12 These small variations between the individual spectra, obtained with increasing dilutions (Figure 1), were due to the slight geometric variations between individual NMR tubes. Moreover, there is no certain way of ensuring that the NMR tube holder will be placed in the exactly same horizontal position when repeatedly being lowered into the magnet. These tiny variations could have been removed if individual magnetic ®eld homogenizations (eg shimming) of each sample had been performed. However, this was not possible here, since the homeopathic solutions did not contain any heavy water (D2O) or other substance containing deuterium which could be used for shimming the lock signal (see Materials and methods). Addition of D2O to the homeopathic solutions would have
NMR spectra of Sulphur and Betula S Aabel et al
changed the test objects and could therefore have been criticized. Similar results were obtained at 500 MHz. No signi®cant differences between various Sulphur potencies and the control solution were observed (Figure 2). Furthermore, in the experiment with the remedy Betula 30c (liquid form) and its two controls (unsuccussed Betula and deionized water) at 500 MHz seemingly identical one dimensional proton spectra (1H NMR) were obtained (Figure 3, enlargement). The T1 values of the D10, D15, D23 and D30 potentiated Sulphur solutions and deionized water were within experimental errors equivalent at 24 C and 20 MHz (Table 1). The statistical analysis con®rmed this conclusion (P 0.83). A shorter T1 was obtained for the D4 Sulphur dilution. The T1 values of the Betula alba 30c preparation (as globules) and the appropriate placebo control were
identical at 37 C (Table 1). At 24 C, the T1 values, though being slightly different, were considered equivalent taking into account the uncertainty of the relaxation measurements (3%).
17
Discussion The one dimensional proton (1H NMR) spectra showed in Figures 1 and 2 were obtained to test the reproducibility of ®ndings indicating that homeopathic solutions may give NMR spectra different from those of control samples (ethanol=water).12 In our experimental set-up we tried to follow the procedure described in WeingaÈrtner,12 although few experimental details were provided. We were not able to reproduce the change in 1H NMR line widths, between individual dilution=succussion stages as
Figure 1 300 MHz 1H NMR spectra of a Sulphur preparation in ethanol=water at varying succussion=dilution stages, from bottom and up (1=106, 1=1012, 1=1024, 1=1030). Top row shows the control solution (ethanol 87%). AU arbitrary units. British Homeopathic Journal
NMR spectra of Sulphur and Betula S Aabel et al
18
Figure 2 500 MHz 1H NMR spectra of Sulphur preparation in ethanol=water at varying succussion=dilution stages, from bottom and up (1=104, 1=1013, 1=1023, 1=1030). Top row shows the control solution (ethanol 87%).
Figure 3 500 MHz 1H NMR spectra of Betula 30c in water (succussed and diluted) (top trace) and the same solution of Betula unsuccussed (only diluted) (bottom). Insert: enlarged part of the marked parts of the spectra, showing the signals from the ethanol. British Homeopathic Journal
NMR spectra of Sulphur and Betula S Aabel et al Table 1 T1 water proton relaxation times of placebo, Betula 30c, water and Sulphur preparations at 20 MHz Preparation Placebo Betula 30c Water Sulphur D4 Sulphur D10 Sulphur D15 Sulphur D23 Sulphur D30
T1 (ms), 24 C 2680 2590 2791 2283 2830 2795 2816 2819
(2622, (2578, (2768, (2274, (2726, (2721, (2731, (2754,
2738) 2602) 2815) 2292) 2933) 2870) 2901) 2883)
T1 (ms), 37 C 3358 (3334, 3381) 3358 (3331, 3385)
Mean values of ®ve replicate analyses are given, with their 95% con®dence intervals in parenthesis, at two different temperatures for placebo and Betula 30c and at one temperature for water and different potencies of Sulphur. Placebo and Betula were sucrose tablets dissolved in water.
reported by WeingaÈrtner.12 One possible explanation for WeingaÈrtners' results is that NMR tubes with markedly different geometric or other properties were employed in his study. Obviously, there is no way of falsifying this assumption now. Another possibility explaining the discrepancy might be that the control solution of WeingaÈrtner12 were not treated in exactly the same way as the homeopathic solutions during the dilution=succussion stages, resulting in different levels of dissolved gases (eg oxygen or carbon dioxide or both). The time elapsing after each dilution=succussion stage before the acquisition of the NMR data might also be important. The dilution=succussion involves heavy shaking and dissolution of air in the samples. It is obvious that if the time period from the dilution= succussion to the NMR measurement is not kept constant, different levels of dissolved gases will be present. Also, the number and sizes of gas bubbles created in the NMR tubes will differ. In the present study the NMR experiments were performed immediately after dilution=succussion to ensure identical conditions of measurement. In particular, an enhanced level of oxygen can induce shorter relaxation times and broader NMR signals.18 Furthermore, an increased level of carbon dioxide stemming from the shaking=succussion procedure, will increase the acidity of the homeopathic solution, which might change the appearance and positions of the resonances. Small changes in pH can change the appearance of hydroxyl (OH) resonances drastically.19,20 Such artifacts may have played a role, since WeingaÈrtner12 discussed the relevance of the different resonances of H2O and OH in ethanol in the control sample versus the combined signal observed in the homeopathic sample. In our case all samples were given the same physical treatment and exposed to the same atmospheric conditions to avoid such false results. However, it was beyond the scope of this study to investigate this possibility in more detail; our endeavor was to investigate if it was possible at all to observe any real difference between homeopathic solutions and control samples.
It is well known that the T1 relaxation times are affected by the temperature and acidity of the medium.21,22 The T1 values of the various preparations investigated here were in accordance with this, being longer at 37 C than at 24 C, due to the lower microviscosity at the higher temperature (Table 1). The slightly shorter T1 value for the potentiated Sulphur D4 compared to higher potencies was probably due to the higher microviscosity within the sample matrix, caused by the high concentration of dissolved material. The lack of any statistically signi®cant difference in T1 between higher potencies of Sulphur and control medium suggests no change in the water proton relaxation properties with increasing potency. Furthermore, there was no difference in relaxation times between potentised medicines (Betula alba 30c globules) and controls (pure sucrose globules). The results reported here thus contrast those reported by Demangeat et al.13 In that study, a lengthened T1 water proton relaxation time was observed at 4 MHz for high saline dilutions of a silica=lactose mixture. Efforts to elucidate the origin of this unexpected ®nding were made but no explanation could be derived from their published results. It is unlikely that different results would have been obtained in this study if relaxometric measurements had been performed at a lower ®eld strength (such as 4 MHz) rather than 20 MHz. In conclusion, there is no experimental evidence that homeopathic remedies make any kind of imprint on their solvent, which can be detected with nuclear magnetic resonance.
19
Acknowledgements This study was ®nanced by the Research Council of Norway. The homeopathic remedies were delivered from DCG, Gothenburg, Sweden. Professor Haakon Benestad provided valuable advice on the manuscript.
References 1 Vallance AK. Can biological activity be maintained at ultrahigh dilution? An overview of homeopathy, evidence and Bayesian Philosophy. J Alternative and Compl Med 1998; 4: 49 ± 76. 2 Linde K, Melchart D. Der Wirksamkeitsnachweis der Homoeopathie Ð eine unendliche Geschichte? Deutsche Apotheker Zeitung 1996; 32: 17 ± 22. 3 Linde K, Clausius N, Ramirez G, et al. Are the clinical effects of homeopathy placebo effects? A meta-analysis of placebo controlled trials. Lancet 1997; 350: 834 ± 843. 4 Walach H. Magic of signs: a nonlocal interpretation of homeopathy. J Scient Exploration. 1999; 13: 291 ± 315. 5 HomoÈopatische Artzneibuch 1. Ausgabe, 1978 3. Nachtrag, 1985, p 405 ± 407. 6 Freeman R. Spin Choreography. Oxford: Spektrum Academic Publishers, 1997. British Homeopathic Journal
NMR spectra of Sulphur and Betula S Aabel et al
20
7 Ernst RR, Bodenhausen G, Wokaun A. Principles of Nuclear Magnetic Resonance in One and Two Dimensions. Oxford: Oxford Science Publications, 1999. 8 GuÈnther H. NMR Spectroscopy. Basic Principles, Concepts, and Applications in Chemistry. Second Edition. New York: Wiley, 1996. 9 McCarthy MJ, Cheung MK. Magnetic Resonance Imaging: Annual Reports on NMR Spectroscopy, London: Academic Press, vol 31, 1995. 10 Young TM. Nuclear Magnetic Resonance studies of succussed solutions. A preliminary report. J American Inst Homeopathy 1975; 68: 8 ± 16. 11 Sacks AD. Nuclear Magnetic Resonance spectroscopy of homeopathic remedies. J Holistic Med 1983; 5: 172 ± 177. 12 WeingaÈrtner O. NMR-features that relate to homeopathic sulphur-potencies. Berlin J Res in Homeopathy 1990; 1: 61 ± 68. 13 Demangeat JL, Demangeat C. Modi®cations des temps de relaxation RMN aÁ 4 Mhz des protons du solvant dans les treÁs hautes dilutions salines de silice=lactose. J Med Nucl Biophysics 1992; 2: 135 ± 145. 14 Eisenberg DM, Davis RB, Ettner S, et al. Trends in alternative medicine use in the United States, 1990 ± 1997. Results of a follow-up national survey. JAMA 1998; 280: 1569 ± 1575.
British Homeopathic Journal
15 Aabel S. Lñrum E, Dùlvik S, Djupesland P. Is homeopathic `immunotherapy' effective? A double-blind, placebo-controlled trial with the isopathic remedy Betula 30c. Br Hom J 2000; 89: 161 ± 168. 16 Braun S, Kalinowski HO, Berger S. 150 and More Basic NMR Experiments. Weinhein: Wiley-VCH, 1998, p 155. 17 Freeman R. Spin Choreography. Oxford: Spektrum Academic Publishers, 1997, p 345. 18 Friebolin H. Basic One- and Two-Dimensional NMR Spectroscopy. Weinhein: VCH Verlagsgesellschaft, 1998, p 178. 19 Friebolin H. Basic One- and Two-Dimensional NMR Spectroscopy. Weinhein: VCH Verlagsgesellschaft, 1998, pp 59 ± 60. 20 Sanders JKM, Hunter BK. Modern NMR Spectroscopy. Oxford: Oxford University Press, 1987, pp 215 ± 216, pp 222 ± 224. 21 Knispel RR, Pintar MM. Temperature dependence of the proton exchange time in pure water by NMR. Chemical Physics Letters 1975; 32: 238 ± 240. 22 Glick RE, Tewari KC. Proton nuclear Magnetic Relaxation studies on water: the rates of acid and base-catalyzed proton exchange. J Chem Phys 1966; 44: 546 ± 547.
ORIGINAL PAPER
Nuclear magnetic resonance (NMR) studies of homeopathic solutions S Aabel1*, S Fossheim2 and F Rise3 Institute of General Practice and Community Medicine, Department of General Practice, University of Oslo, Oslo, Norway; 2Department of Pharmacy, University of Oslo, Oslo, Norway; and 3Department of Chemistry, University of Oslo, Oslo, Norway 1
The ef®cacy of homeopathy is controversial. Nuclear magnetic resonance (NMR) has been used to study homeopathic solutions, showing provocative results. We examined the reproducibility of one of the allegedly positive studies. 1 H NMR spectra were recorded for Sulphur D4, diluted and succussed up to D30 (called potentization) at two different frequencies (300 and 500 MHz). The Sulphur solution had been potentiated according to homeopathic principles with deionized water and alcohol. Water proton T1 relaxation measurements were performed also at 20 MHz for the different potentiated Sulphur solutions. Furthermore, the homeopathic remedy Betula alba 30c (birch pollen extract) and appropriate control solution (deionized water, unsuccussed solutions and placebo globules) were measured analogously, both with frequencies giving spectra and T1 relaxometry. The Sulphur remedies showed identical one dimensional proton spectra (1H NMR) at 300 and 500 MHz, regardless of dilution=succussion stage, from D4 to D30. Furthermore, Betula 30c as a potentiated solution and its controls (ethanol dilutions and Betula diluted but not succussed) showed identical spectra. Nor were there any statistically signi®cant differences in longitudinal (T1) relaxation times between deionized water and Sulphur D10 to D30 preparations. The shorter T1 of the Sulphur D4 preparation could be ascribed to the higher microviscosity within the sample matrix caused by the high concentration of dissolved material. Also, the T1 values of the Betula alba 30c preparation (in globular form) and control placebo globules were identical. In conclusion, published results from NMR research on homeopathy indicating differences between homeopathic solutions and control samples could not be reproduced. British Homeopathic Journal (2001) 90, 14±20. Keywords: homeopathy; nuclear magnetic resonance; Betula 30c; Sulphur D4 ± D30
Introduction The ef®cacy of homeopathy is controversial1 This may be due to the fact that several studies, both clinical and laboratory, apparently have shown positive results without providing plausible mechanistic explanations.2 ± 4 The most controversial point is the very high dilution of the homeopathic remedies. The solutions are in fact produced by a combination of *Correspondence: S Aabel, Institute of General Practice and Community Medicine, Department of General Practice, University of Oslo, PO Box 1130 Blindern, N-0317 Oslo, Norway. Received 10 April 2000; revised 27 June 2000; accepted 22 September 2000
dilution and succussion (called potentization), according to certain speci®c procedures.5 Accordingly, the remedies prescribed by a homeopath may well be diluted so much that there are no molecules of the remedy left in the solution. In theory, the probability of detecting even one molecule of the selected agent is zero or close to zero in homeopathic solutions when the solution is diluted more than 1024 times (ie D24). Several nuclear magnetic resonance (NMR) spectroscopic methods are well suited to study certain kinetic and structural properties of water, both in vitro and in tissue.6 ± 9 Therefore, NMR has been used to examine various homeopathic solutions.10 ± 13 Some results are certainly provocative, reporting a structural difference between the one dimensional 1H NMR
NMR spectra of Sulphur and Betula S Aabel et al
spectra of homeopathic solutions and the unsuccussed solvent of the remedy, ie 87% (v=v) ethanol.12 Since homeopathy is quite popular in the western world,14 and NMR investigations of homeopathic solutions are much easier to perform and assess than clinical studies, we decided to examine the reproducibility of one of the allegedly positive studies.12 Moreover, since a clinical trial on isopathic therapy has earlier been conducted, where the remedy was either a high potency of a birch pollen preparation (Betula 30c) or a placebo preparation,15 samples of this remedy were also investigated by NMR.
Materials and methods Chemicals
Sulphur D4 (diluted 1=104) (dilutions prepared by DCG, Drogesentralen, Gothenburg. Sulphur Merck No 4262458 DAB 7 1992), de-ionized water, 96% (v=v) and 87% (v=v) ethanol, diluted from 96% (v=v) ethanol (Arcus Products, Oslo) with deionized water were used. D2O (deuterium oxide) 99.96% D, was purchased from Cambridge Isotope Labs (CIL Andover, MA). Betula alba 30c tablets (sucrose globules impregnated with a Betula alba 30c solution), placebo tablets (ie sucrose globules), Betula alba 30c dilution (succussed and diluted) and Betula alba (diluted but not succussed) were employed. All Betula alba preparations were made from birch pollen, as described in HomoÈopatische Artzneibuch5 and delivered from DCG. Preparation of homeopathic dilutions
Sulphur. The homeopathic medicine delivered from the producer consisted of pulverized Sulphur with added alcohol and water in a dilution of 1=104 (D4). This solution was diluted with 87% ethanol and further diluted and succussed down to 1=1030 (D30) according to standard homeopathic procedures.5 Brie¯y, 26 unused, capped (plastic) glass ¯asks were washed once with deionized water and ¯ushed with 96% ethanol, dried at 160 C for 1 h, cooled to room temperature and then marked with dilution numbers. Of the initial Sulphur solution, (1=104) 5 ml was pipetted to a glass ¯ask to which was also added 45 ml of solvent (87% ethanol). The ¯ask was closed and vigorously succussed 10 times against an unyielding surface according to homeopathic principles. Then, the cleaned pipette (50% v=v 96% ethanol in deionized water) was reused to transfer an unspeci®ed amount of the solution to a new NMR tube (vide supra) for cleaning purposes (twice), before 2 ml (for relaxometry at 20 MHz), 0.7 ml or 0.4 ml (1H NMR spectroscopy at 300 or 500 MHz, respectively) were transferred from the ¯ask in to the NMR tube. The NMR tubes were immediately sealed with standard plastic caps to avoid further exposure to the surrounding atmosphere. Analogously, the dilution process was
continued, so that a 5 ml sample from the 1=105 solution was transferred to the next container. To this solution, 45 ml of solvent was also added and the mixture was again succussed 10 times. This procedure was repeated until a solution diluted to 1=1030 had been prepared. The dilution=succussion procedures were performed in the air-conditioned NMR laboratories to ensure constant temperature conditions. Every new dilution was assessed with one dimensional 1H NMR measurements to see if a change in the appearance (line width and intensity) of the ethanol signals could be detected. Also, the longitudinal (T1) relaxation times of these potentiated solutions were measured at 20 MHz and 37 C.
15
Betula alba 30c, Betula diluted, placebo and ethanol. The Betula alba 30c remedy, prepared by DCG (as used in experiments reported elsewhere5) was both in liquid and globule from. Betula alba 30c is made from birch pollen. The liquid preparation was succussed and diluted 30 times the solvent being 8% (v=v) ethanol in distilled water. The Betula alba 30c solution was examined at 500 MHz (1H NMR spectroscopy). Betula diluted, a liquid preparation that had only been diluted (with no succussion) 30 times was also examined at 500 MHz. For the T1 relaxation experiments globules of Betula alba 30c were used, meaning that the liquid preparation had impregnated sucrose globules. These Betula alba 30c globules were dissolved in deionzied water prior to the measurements. Placebo globules prepared of pure sucrose were dissolved in deionzied water and measured just like the Betula alba 30c preparation. NMR instruments and ancillary equipment
The one dimensional 1H NMR experiments were performed with Varian XL300 instrument (Palo Alto, CA) (manual) (tuned to 299.944 MHz) equipped with a standard 5 mm 1H=broad band switchable probe and with a Bruker Avance DRX 500 instrument (FaÈllanden, Switzerland) (tuned to 500.13 MHz) that had an inverse triple resonance probe (TXI). The longitudinal (T1) relaxation measurements were performed on a 20 MHz instrument (Minispec PC 120-b Rheinstetten, Germany). The NMR sample tubes used at 300 and 500 MHz were Wilmad Economy (WG5-ECONOMY 5 mm o.d.) tubes (Wilmad Glass Co, Buena, NJ). The samples were investigated with GRtubes (dimension 18010 mm) on the Minispec instrument. NMR methods, computers and software
One dimensional proton spectra (1H NMR) were obtained with the standard pulse program STD1H, delivered with the Varian XL300 spectrometer and acquired and processed with the Varian Magical 6.2B British Homeopathic Journal
NMR spectra of Sulphur and Betula S Aabel et al
16
program in the program in the proprietary Varian computer, equipped with a Motorola 68000 processor. An acquisition time of 3.75 s, pulse width 1 ms (PW 90 25.2 ms), 16 scans (NS) per sample and a delay of 0 s between the scans and the receiver gain, reduced to minimum (manually), were used with the XL 300 instrument. Furthermore, a line broadening (LB) factor of 0.3 Hz was used in the processing. The acquisitions were performed at 20 C on non spinning samples. One dimensional proton spectra (1H NMR) were also acquired with a Bruker Avance DRX 500 instrument and processed on Silicon Graphics Indy computers with the program Xwinnmr (versions 1.3 and 2.1) employing the zg30 (30 degree pulse angle) pulse program. A delay (D1) of one second between each pulse was employed, a total of 16 scans were used for each acquisition, with an acquisition time of 3.17 s used for each scan. The acquisitions were performed on non spinning samples at 25 C, without sweeping the lock channel and without deuterium lock, since no D2O (heavy water) was introduced into the homeopathic samples. A minimum receiver gain of one was used due to the intense 1H signals. A line broadening factor of 1 Hz were employed when processing the 1H one dimensional spectra. The magnetic ®eld homogeneity of the 300 MHz and 500 MHz instruments were optimized (shimmed) on the lock signal with a sample where one part of heavy water (D2O) was added to two parts of 87% ethanol. The same shim parameters were used for each session of successive dilutions. The lock functionalities were then disabled, and the receiver gain on the 1H channels were reduced to a minimum before the actual measurements were performed. No corrections due to variations of the geometric or other properties of the individual NMR tubes were performed between each dilution and each replicate.
1
H NMR spectra of ethanol with different Sulphur dilutions at 300 MHz and 500 MHz
Each potentiated dilution was measured at 300 MHz and every ®fth potentiated dilution was monitored at 500 MHz with one of the NMR instruments as described above. Each dilution investigated was assessed with respect to the intensities and line widths of the 1H NMR signals of ethanol.
1
H NMR spectra of ethanol=water in the presence of different Betula dilutions
The experiments were performed as the Sulphur experiments described above, except that spectra were only recorded at 500 MHz.
British Homeopathic Journal
Longitudinal (T1) relaxation time constants for Sulphur dilutions, Betula 30c globules and placebo globules at 20 MHz
The relaxometric study was performed at 20 MHz to avoid the problems encountered with radiation damping, using high quality probes at higher ®elds (300 MHz and 500 MHz).17 The water proton T1 relaxation times of the Betula alba 30c and placebo preparations were recorded at both 24 and 37 C, with a relaxation delay of 15 and 20 s, respectively. The T1 values of the different Sulphur potentiated dilutions and deionized water were only recorded at 24 C. For each sample, the T1 relaxation time was measured ®ve times. Prior to any measurement, the spectrometer was calibrated by adjusting the ®eld homogeneity, the 90 and 180 pulse widths. The T1 water proton relaxation times were determined from 15 data points, using an inversion recovery pulse sequence.16 As a control of the spectrometer's reproducibility, the T1 relaxation time of an NMR standard (1 mM solution of gadolinium diethylenetriamine penta-acetic acid, diluted Magnevist1, Schering AG, Berlin, Germany) was measured at 37 C. The precision of the instrument was set at 3%. The T1 relaxation data are given as means with their 95% con®dence intervals. One-way analysis of variance was used to test for statistical differences. A probability (P) value less than 0.05 was considered statistically signi®cant.
Results All diluted and succussed solutions of the Sulphur showed very similar one dimensional proton spectra (1H NMR) at 300 MHz, regardless of dilution=succussion stage as did the control dilution without the homeopathic ingredient. In other words, the CH2 and CH3 water (H2O) and ethanol hydroxy (OH) resonances did not change appearance in a physiochemically signi®cant way between each dilution=succussion stage (Figure 1). The minimal differences which we found, were far less than the changes reported by WeingaÈrtner.12 These small variations between the individual spectra, obtained with increasing dilutions (Figure 1), were due to the slight geometric variations between individual NMR tubes. Moreover, there is no certain way of ensuring that the NMR tube holder will be placed in the exactly same horizontal position when repeatedly being lowered into the magnet. These tiny variations could have been removed if individual magnetic ®eld homogenizations (eg shimming) of each sample had been performed. However, this was not possible here, since the homeopathic solutions did not contain any heavy water (D2O) or other substance containing deuterium which could be used for shimming the lock signal (see Materials and methods). Addition of D2O to the homeopathic solutions would have
NMR spectra of Sulphur and Betula S Aabel et al
changed the test objects and could therefore have been criticized. Similar results were obtained at 500 MHz. No signi®cant differences between various Sulphur potencies and the control solution were observed (Figure 2). Furthermore, in the experiment with the remedy Betula 30c (liquid form) and its two controls (unsuccussed Betula and deionized water) at 500 MHz seemingly identical one dimensional proton spectra (1H NMR) were obtained (Figure 3, enlargement). The T1 values of the D10, D15, D23 and D30 potentiated Sulphur solutions and deionized water were within experimental errors equivalent at 24 C and 20 MHz (Table 1). The statistical analysis con®rmed this conclusion (P 0.83). A shorter T1 was obtained for the D4 Sulphur dilution. The T1 values of the Betula alba 30c preparation (as globules) and the appropriate placebo control were
identical at 37 C (Table 1). At 24 C, the T1 values, though being slightly different, were considered equivalent taking into account the uncertainty of the relaxation measurements (3%).
17
Discussion The one dimensional proton (1H NMR) spectra showed in Figures 1 and 2 were obtained to test the reproducibility of ®ndings indicating that homeopathic solutions may give NMR spectra different from those of control samples (ethanol=water).12 In our experimental set-up we tried to follow the procedure described in WeingaÈrtner,12 although few experimental details were provided. We were not able to reproduce the change in 1H NMR line widths, between individual dilution=succussion stages as
Figure 1 300 MHz 1H NMR spectra of a Sulphur preparation in ethanol=water at varying succussion=dilution stages, from bottom and up (1=106, 1=1012, 1=1024, 1=1030). Top row shows the control solution (ethanol 87%). AU arbitrary units. British Homeopathic Journal
NMR spectra of Sulphur and Betula S Aabel et al
18
Figure 2 500 MHz 1H NMR spectra of Sulphur preparation in ethanol=water at varying succussion=dilution stages, from bottom and up (1=104, 1=1013, 1=1023, 1=1030). Top row shows the control solution (ethanol 87%).
Figure 3 500 MHz 1H NMR spectra of Betula 30c in water (succussed and diluted) (top trace) and the same solution of Betula unsuccussed (only diluted) (bottom). Insert: enlarged part of the marked parts of the spectra, showing the signals from the ethanol. British Homeopathic Journal
NMR spectra of Sulphur and Betula S Aabel et al Table 1 T1 water proton relaxation times of placebo, Betula 30c, water and Sulphur preparations at 20 MHz Preparation Placebo Betula 30c Water Sulphur D4 Sulphur D10 Sulphur D15 Sulphur D23 Sulphur D30
T1 (ms), 24 C 2680 2590 2791 2283 2830 2795 2816 2819
(2622, (2578, (2768, (2274, (2726, (2721, (2731, (2754,
2738) 2602) 2815) 2292) 2933) 2870) 2901) 2883)
T1 (ms), 37 C 3358 (3334, 3381) 3358 (3331, 3385)
Mean values of ®ve replicate analyses are given, with their 95% con®dence intervals in parenthesis, at two different temperatures for placebo and Betula 30c and at one temperature for water and different potencies of Sulphur. Placebo and Betula were sucrose tablets dissolved in water.
reported by WeingaÈrtner.12 One possible explanation for WeingaÈrtners' results is that NMR tubes with markedly different geometric or other properties were employed in his study. Obviously, there is no way of falsifying this assumption now. Another possibility explaining the discrepancy might be that the control solution of WeingaÈrtner12 were not treated in exactly the same way as the homeopathic solutions during the dilution=succussion stages, resulting in different levels of dissolved gases (eg oxygen or carbon dioxide or both). The time elapsing after each dilution=succussion stage before the acquisition of the NMR data might also be important. The dilution=succussion involves heavy shaking and dissolution of air in the samples. It is obvious that if the time period from the dilution= succussion to the NMR measurement is not kept constant, different levels of dissolved gases will be present. Also, the number and sizes of gas bubbles created in the NMR tubes will differ. In the present study the NMR experiments were performed immediately after dilution=succussion to ensure identical conditions of measurement. In particular, an enhanced level of oxygen can induce shorter relaxation times and broader NMR signals.18 Furthermore, an increased level of carbon dioxide stemming from the shaking=succussion procedure, will increase the acidity of the homeopathic solution, which might change the appearance and positions of the resonances. Small changes in pH can change the appearance of hydroxyl (OH) resonances drastically.19,20 Such artifacts may have played a role, since WeingaÈrtner12 discussed the relevance of the different resonances of H2O and OH in ethanol in the control sample versus the combined signal observed in the homeopathic sample. In our case all samples were given the same physical treatment and exposed to the same atmospheric conditions to avoid such false results. However, it was beyond the scope of this study to investigate this possibility in more detail; our endeavor was to investigate if it was possible at all to observe any real difference between homeopathic solutions and control samples.
It is well known that the T1 relaxation times are affected by the temperature and acidity of the medium.21,22 The T1 values of the various preparations investigated here were in accordance with this, being longer at 37 C than at 24 C, due to the lower microviscosity at the higher temperature (Table 1). The slightly shorter T1 value for the potentiated Sulphur D4 compared to higher potencies was probably due to the higher microviscosity within the sample matrix, caused by the high concentration of dissolved material. The lack of any statistically signi®cant difference in T1 between higher potencies of Sulphur and control medium suggests no change in the water proton relaxation properties with increasing potency. Furthermore, there was no difference in relaxation times between potentised medicines (Betula alba 30c globules) and controls (pure sucrose globules). The results reported here thus contrast those reported by Demangeat et al.13 In that study, a lengthened T1 water proton relaxation time was observed at 4 MHz for high saline dilutions of a silica=lactose mixture. Efforts to elucidate the origin of this unexpected ®nding were made but no explanation could be derived from their published results. It is unlikely that different results would have been obtained in this study if relaxometric measurements had been performed at a lower ®eld strength (such as 4 MHz) rather than 20 MHz. In conclusion, there is no experimental evidence that homeopathic remedies make any kind of imprint on their solvent, which can be detected with nuclear magnetic resonance.
19
Acknowledgements This study was ®nanced by the Research Council of Norway. The homeopathic remedies were delivered from DCG, Gothenburg, Sweden. Professor Haakon Benestad provided valuable advice on the manuscript.
References 1 Vallance AK. Can biological activity be maintained at ultrahigh dilution? An overview of homeopathy, evidence and Bayesian Philosophy. J Alternative and Compl Med 1998; 4: 49 ± 76. 2 Linde K, Melchart D. Der Wirksamkeitsnachweis der Homoeopathie Ð eine unendliche Geschichte? Deutsche Apotheker Zeitung 1996; 32: 17 ± 22. 3 Linde K, Clausius N, Ramirez G, et al. Are the clinical effects of homeopathy placebo effects? A meta-analysis of placebo controlled trials. Lancet 1997; 350: 834 ± 843. 4 Walach H. Magic of signs: a nonlocal interpretation of homeopathy. J Scient Exploration. 1999; 13: 291 ± 315. 5 HomoÈopatische Artzneibuch 1. Ausgabe, 1978 3. Nachtrag, 1985, p 405 ± 407. 6 Freeman R. Spin Choreography. Oxford: Spektrum Academic Publishers, 1997. British Homeopathic Journal
NMR spectra of Sulphur and Betula S Aabel et al
20
7 Ernst RR, Bodenhausen G, Wokaun A. Principles of Nuclear Magnetic Resonance in One and Two Dimensions. Oxford: Oxford Science Publications, 1999. 8 GuÈnther H. NMR Spectroscopy. Basic Principles, Concepts, and Applications in Chemistry. Second Edition. New York: Wiley, 1996. 9 McCarthy MJ, Cheung MK. Magnetic Resonance Imaging: Annual Reports on NMR Spectroscopy, London: Academic Press, vol 31, 1995. 10 Young TM. Nuclear Magnetic Resonance studies of succussed solutions. A preliminary report. J American Inst Homeopathy 1975; 68: 8 ± 16. 11 Sacks AD. Nuclear Magnetic Resonance spectroscopy of homeopathic remedies. J Holistic Med 1983; 5: 172 ± 177. 12 WeingaÈrtner O. NMR-features that relate to homeopathic sulphur-potencies. Berlin J Res in Homeopathy 1990; 1: 61 ± 68. 13 Demangeat JL, Demangeat C. Modi®cations des temps de relaxation RMN aÁ 4 Mhz des protons du solvant dans les treÁs hautes dilutions salines de silice=lactose. J Med Nucl Biophysics 1992; 2: 135 ± 145. 14 Eisenberg DM, Davis RB, Ettner S, et al. Trends in alternative medicine use in the United States, 1990 ± 1997. Results of a follow-up national survey. JAMA 1998; 280: 1569 ± 1575.
British Homeopathic Journal
15 Aabel S. Lñrum E, Dùlvik S, Djupesland P. Is homeopathic `immunotherapy' effective? A double-blind, placebo-controlled trial with the isopathic remedy Betula 30c. Br Hom J 2000; 89: 161 ± 168. 16 Braun S, Kalinowski HO, Berger S. 150 and More Basic NMR Experiments. Weinhein: Wiley-VCH, 1998, p 155. 17 Freeman R. Spin Choreography. Oxford: Spektrum Academic Publishers, 1997, p 345. 18 Friebolin H. Basic One- and Two-Dimensional NMR Spectroscopy. Weinhein: VCH Verlagsgesellschaft, 1998, p 178. 19 Friebolin H. Basic One- and Two-Dimensional NMR Spectroscopy. Weinhein: VCH Verlagsgesellschaft, 1998, pp 59 ± 60. 20 Sanders JKM, Hunter BK. Modern NMR Spectroscopy. Oxford: Oxford University Press, 1987, pp 215 ± 216, pp 222 ± 224. 21 Knispel RR, Pintar MM. Temperature dependence of the proton exchange time in pure water by NMR. Chemical Physics Letters 1975; 32: 238 ± 240. 22 Glick RE, Tewari KC. Proton nuclear Magnetic Relaxation studies on water: the rates of acid and base-catalyzed proton exchange. J Chem Phys 1966; 44: 546 ± 547.