Reduction in Serum Leptin and IGF-1 but Preserved T-Lymphocyte Numbers and Activation After a Ketogenic Diet in Rheumatoid Arthritis Patients

Home , Ketogenic diet

Reduction in serum leptin and IGF-1 but preserved T-lymphocyte numbers and activation after a ketogenic diet in rheumatoid arthritis patients

D.A. Fraser1,4, J. Thoen1, S. Bondhus1, M. Haugen1, J.E. Reseland2, O. Djøseland3, Ø. Førre1, J. Kjeldsen-Kragh4

1Centre for Rheumatic Diseases, The National Hospital; 2Institute for Nutrition Research, University of Oslo; 3Department of Endocrinology, The National Hospital; 4Department of Immunology and Transfusion Medicine, Ullevaal University Hospital, Oslo, Norway.

Abstract Objective To assess the clinical, immunological and hormonal effects of carbohydrate restriction in rheumatoid arthritis (RA) patients via the provision of a ketogenic diet. Methods Thirteen RA patients with active disease consumed a ketogenic diet for 7 days, providing the estimated require- ments for energy and protein whilst restricting their carbohydrate intake to < 40 g/day. This was followed by a 2- week re-feeding period. Clinical and laboratory evaluations were carried out on days 0, 7 and 21. Changes in serum glucose, -hydroxybutyrate ( -HB), leptin, insulin-like growth factor-1 (IGF-1) and cortisol were also measured at these time points. To study CD4+ and CD8+ lymphocyte responses, mitogen stimulated T-cell activation was assessed in heparinised whole blood via flow-cytometric analysis of CD69 expression. Results After the 7-day ketogenic diet, there were significant increases in serum -HB and cortisol, and significant decreases in body weight, the total lymphocyte count, serum leptin, IGF-1 and glucose. However, with the exception of morning stiffness, there were no significant changes in any of the clinical or laboratory measures of disease activity, or in early T-lymphocyte activation and the absolute numbers of CD4+ and CD8+ cells. Conclusion In RA patients several of the metabolic and hormonal responses to a ketogenic diet, such as a fall in serum IGF-1 and leptin, resemble those which occur in response to acute starvation. However, the clinical and immunological changes which occur in response to acute starvation do not take place with a ketogenic diet and thus may be dependent upon energy and/or protein restriction. Key words Rheumatoid arthritis, ketogenic diet, leptin, IGF-1, fasting, immune.

Clinical and Experimental Rheumatology 2000; 18: 209-214. Ketogenic diet and rheumatoid arthritis / D.A. Fraser et al.

David A. Fraser, MSc; Jørn Thoen, MD, Introduction levels, suggesting that the reduced im- PhD; Sissel Bondhus, BSc; Margaretha Beneficial effects of acute starvation mune function associated with starvation Haugen, PhD; Janne E. Reseland, PhD; upon both clinical and laboratory vari- may be leptin-mediated (18). We found Ole Djøseland, PhD; Øystein Førre, MD, ables reflecting disease activity in rheu- that serum leptin concentrations were PhD; Jens Kjeldsen-Kragh, MD, PhD. matoid arthritis (RA) have been docu- reduced by 67% after fasting in RA pa- Please address correspondence and reprint mented (1-3). We have recently shown tients (4). As leptin enhances both naive requests to: David A. Fraser, Department in RA patients that a 7-day fast, in addi- CD4+ lymphocyte proliferative respon- of Immunology and Transfusion Medicine, tion to causing significant decreases in ses and Th1 cytokine production (18), a Ullevaal University Hospital, N-0407, Oslo, Norway. ESR, CRP, the tender joint count and fall in serum leptin might explain both E-mail: [email protected] morning stiffness, also decreases CD4+ the increase in IL-4 production and the lymphocyte activation and numbers decrease in CD4+ T-cell activation pro- © Copyright CLINICAL AND whilst increasing IL-4 production from duction in mitogen stimulated peripheral EXPERIMENTAL RHEUMATOLOGY 2000. peripheral blood cells (4). Fasting has blood cells we recently observed (4). also been shown to decrease mitogen- We speculated that the immunosuppres- and antigen-induced lymphocyte prolif- sive and clinical effects of acute starva- erative responses (5) and to suppress tion in RA patients might be dependent interleukin-2 production in healthy sub- upon hormonal adaptations to dietary jects (6). carbohydrate restriction rather than upon Although it is well established that chro- energy and/or protein deprivation. We nic sub-optimal protein and calorie in- have therefore assessed the effects of a takes have a negative effect upon cell- 7-day ketogenic diet (2000-2500 kcal/ mediated immune function (7, 8), the day primarily as lipids, 0.8 g protein/kg mechanisms underlying immunological body weight/day, < 40 g carbohydrate/ changes associated with acute starvation day) upon laboratory and clinical indi- in both healthy subjects and RA patients ces of disease activity in a group of 13 are unclear. The lack of any clinical or RA patients. We simultaneously assessed laboratory effects of an elemental diet changes in T-cell number and function does not support the idea that the avoid- along with serum glucose, leptin, IGF- ance of specific food items is responsi- 1, b-hydroxybutyrate (b-HB) and corti- ble for the clinical effects of fasting in sol levels. RA patients (9, 10). Interestingly, it is known that carbohy- Patients and methods drate restriction mediates many of the Patients metabolic and hormonal responses to Thirteen RA patients attending the out-

fasting (11, 12). Thus, the substitution of patient department at the Centre for dietary carbohydrates with lipids results Rheumatic Diseases, The National Hos- in similar decreases in blood glucose and pital, who satisfied the American Col- insulin and increases in free fatty acids, lege of Rheumatology criteria for RA ketone bodies and whole body lipolysis (19) were recruited. All had active dis- as would be observed during total fast- ease, as defined by the presence of 3 of ing. Other hormonal adaptations to car- the 4 following criteria; more than 3 bohydrate restriction may have impor- swollen joints (28-joint score), more than tant immunomodulatory effects. For ex- 6 tender joints (28-joint score), duration ample cortisol, which has well estab- of morning stiffness > 45 mins, erythro- lished anti-inflammatory effects, plays cyte sedimentation rate (ESR) > 28 mm a counter-regulatory role in response to in the 1st hour. All patients had radio- hypoglycaemia (13) and is increased by logical erosions of grade 3 or less, a body an isoenergetic carbohydrate restricted mass index (BMI) between 22 kg/m2 and diet (14). 28 kg/m2 and had not previously partici- A reduction in blood glucose concentra- pated in any dietary regime incorporat- tions may also mediate acute changes in ing fasting, and were not on any self- or serum concentrations of leptin, the pro- medically-prescribed diets which could tein product of the ob gene (15-17). Re- influence the results. Patients with other cently it has been shown that the immu- serious medical conditions were not in- nosuppressive effects of acute starvation cluded in the trial. There were 12 women in mice can be abolished by preventing and 1 man, with a median age of 37 years the fasting-induced fall in plasma leptin (range 25-69 years) and a median dis-

210 Hypothalamus-pituitary-adrenocortical and -gonadal axis in RA / M. CutoloKetogenic diet and rheumatoid arthritis / D.A.EDITORIAL Fraser et al. ease duration of 4 years (range 0.2 - 20 instructed to consume a normal break- temperature the erythrocytes were lysed years). Twelve patients were in func- fast before the evaluation. Two further by a Multi-Q-Prep (Coulter, Hialeah, FL, tional class II and 1 was in functional clinical examinations were carried out on USA). The samples were then analysed class III (20). Eleven patients were rheu- the last morning of the diet (day 7) and by a FACScan or FACSCalibur (Becton matoid factor positive (IgM). Four pa- 2 weeks later on the completion of the Dickinson) flow cytometer. tients were taking second-line drugs, 4 re-feeding period (day 21). The follow- The lymphocytes were identified on the were taking corticosteroids, 3 were taking clinical variables were measured at forward light scatter versus side light ing cytostatic drugs, 13 were taking baseline, on day 7 and on day 21 by the scatter plot. The percentage of CD69- NSAIDS and 4 were taking painkillers. same examining physician: body weight, positive cells was obtained by subtract- The corticosteroid dosage did not exceed morning stiffness (min), the number of ing the percentage of positive events in 5 mg/day and this dose was stable for at tender joints (28 joint score), and the the isotype control sample from the per- least 4 weeks before study entry and number of swollen joints (28 joint score). centage of positive events in the anti- during the trial. Patients using slow-act- At baseline, on day 7 and on day 21 pe- CD69 antibody-stained sample. Lym- ing anti-rheumatic or cytostatic drugs ripheral blood samples were taken by phocyte subset numbers were quantified were on a stable dosage for at least 3 venipuncture between 9:00 and 10:00 using FlowCount fluorospheres (Coul- months prior to study entry. The dosage am. The ESR, CRP, platelet count and ter) by adding a known amount of fluor- of NSAIDs was stable for 2 weeks prior differential white blood cell count ospheres to each tube immediately prior to inclusion. Patients were instructed to (WBC) were measured immediately at to sample analysis on the flow cytom- maintain a stable dosage of their current the Department of Clinical Chemistry, eter. The fluorospheres were gated on the drugs and no changes in drug type were National Hospital, Oslo. Samples of FL1 and FL2 channels and their num- permitted during the trial. heparinised blood (preservative-free so- bers obtained, and the numbers of CD4+ dium heparin 150 U.S.P. units) were also and CD8+ lymphocytes were similarly Study design collected at all time points for analysis gated and counted on the FL1 and FL3 The study was approved by the regional of lymphocyte activation. Serum sam- channels respectively. The number of scientific ethics committee. The 13 pa- ples were frozen at -70°C for subsequent fluorospheres sampled by the flow cy- tients were contacted by telephone by a analysis of cortisol, leptin, IGF-1, glu- tometer as a proportion of the total dietician to establish if there were any cose and b-HB. amount added was then used to calcu- individual dislikes or preferences regard- late the absolute number of CD4+ and ing the ketogenic diet. The diet consisted Assays CD8+ lymphocytes. of selected vegetables, meat, fish, eggs, T-cells were stimulated by a mixture of Serum cortisol, leptin and IGF-1 were nuts, mayonnaise, olive oil, herbs and 2 monoclonal antibodies (mAbs) against measured using commercially available spices. It was prepared by a dietician and CD2, clone L303.1, IgG2a, and CD2R, radioimmunoassay kits (Orion Diagnos- was varied as much as possible to im- clone L304.1, and IgG1 (Becton Dick- tica, Espoo, Finland; Linco Research, St. prove palatability and thus compliance. inson, San Jose, CA, USA) which when Charles, USA; and Nichols Institute Di- The diet provided between 2000 and used together induce polyclonal T-cell agnostics, CA, USA respectively). Se-

2500 kcal (8.4 - 10.5 MJ) per day depend- activation. T-cells were stimulated or left rum glucose and b-HB were measured ing upon body weight and provided 0.8 resting by the addition or not of 5 ml enzymatically (Boerhinger Mannheim, g protein/kg body weight per day and < CD2/CD2R (1 mg/ml) respectively to 250 Mannheim, Germany; and Sigma, St. 40 g carbohydrate/day. All patients were ml aliquots of freshly isolated heparinised Louis, MO, USA). Normal reference instructed to consume sufficient fluids blood. These mixtures were then incu- values for apparently healthy subjects in the form of water, tea, coffee (with- bated for 4 hrs at 37°C in a humidified (supplied by manufacturer) for cortisol out milk) or other calorie-free drinks dur- atmosphere containing 5% CO2. are 167 - 682 nmol/l (8-10 am) and for ing the diet. After 7 days on the ketogenic After the incubation, aliquots of 50 ml fasting glucose are 3.9 - 5.8 mmol/l. diet, patients followed a lacto-vegetar- from the samples with activated cells and ian diet for a period of 2 weeks. The aliquots of 50 ml from the samples with Statistical analysis lacto-vegetarian diet was presented as an resting cells were incubated for 30 min Within group differences were tested by “experimental diet” and thus, to some with the following fluorochrome conju- the Wilcoxon signed-rank test. Two- degree, acted as a placebo for the keto- gated mAbs: anti-CD69, clone L78, sided p values of < 0.05 were consid- genic diet period. All patients were in- IgG1, PE conjugated (Becton Dickinson) ered as significant. The analyses was structed to continue their normal every- + anti-CD4, clone RPA-T4, IgG1, FITC carried out using STATVIEW 4.1 (Aba- day activities during both the ketogenic conjugated (Serotec, Oxford, England) cus Concepts, Berkeley, CA). diet and the 2-week re-feeding period. + anti-CD8, clone SK1, IgG1, PerCP conjugated (Becton Dickinson). Mouse Results Clinical and laboratory assessment IgG1, clone X40, conjugated either with Clinical response All patients attended a baseline exami- PE, FITC and PerCP (all from Becton As is shown in Table I, there were no nation on the same day as the commence- Dickinson) were used as control antibod- significant decreases in any of the ob- ment of the ketogenic diet. They were ies. After 30 min of incubation at room jective variables of disease activity at any

211 Ketogenic diet and rheumatoid arthritis / D.A. Fraser et al. time point. With regard to the subjective sion on either CD4+ or CD8+ cells at observed in response to fasting (4). Our variables, there was a significant de- any time point (Table III). initial hypothesis was that metabolic and crease in morning stiffness at the con- hormonal adaptations to carbohydrate clusion of the ketogenic diet. Discussion restriction, as opposed to energy restric- In this study we have investigated whe- tion, could be responsible for the ob- Metabolic and hormonal response ther a ketogenic diet would modulate served effects of fasting. If this were true, We found a significant decrease in body immune function and improve disease the provision of a ketogenic diet should weight after the ketogenic diet as com- activity in RA patients, as we recently reproduce the beneficial effects of fast- pared to baseline (Table II). There was a significant increase in serum cortisol levels after the ketogenic diet as compared Table I. Effects of ketogenic diet upon clinical and laboratory variables.1 with baseline. The median b-HB concen- Baseline Day 7 of After 2 weeks tration on day 7 of the ketogenic diet was ketogenic diet of re-feeding significantly increased as compared to baseline, whereas serum glucose concen- Tender joints (28-joint score) NS 12 (6 - 16) 8 (5 - 14) 10 (6 - 16) trations were significantly decreased. At Swollen joints (28-joint score) NS 5 (4 - 12) 7 (4 - 13) 10 (4 - 14) the conclusion of the ketogenic diet, both Morning stiffness (mins) 120 (60 - 120) 60 (30 - 120)2 90 (45 - 120) median serum leptin and IGF-1 levels ESR (mm/h) NS 28 (20 - 48) 28 (16 - 40) 30 (18 - 62) had significantly decreased by almost CRP (mg/l) NS 13 (5 - 61) 19 (9 - 56) 12 (5 - 44) 50%. After re-feeding, only body weight remained significantly reduced as com- 1Median (95% confidence intervals); n = 13. 2 pared to baseline, although this differ- Significantly different from baseline (Wilcoxon signed rank) p < 0.05. NS = no significant changes at any timepoint. ence was negligible. No significant differences were observed for any other variables at day 21 compared to base- Table II. Effects of a 7-day ketogenic diet upon body weight, serum cortisol, glucose, b-HB, line. leptin and IGF-1.1

Total and differential WBC and CD4+ Baseline Day 7 of After 2 weeks and CD8+ counts ketogenic diet of re-feeding There was a small but significant de- 3 3 crease in the absolute lymphocyte count Body weight (kg) 65.9(60-80.4) 63 (58.2-77.6) 64.9 (59-79.2) 2 after the ketogenic diet, but no changes Cortisol (nmol/l) 325(272-475) 371(320-577) 319 (269-439) 3 in the total WBC or CD4+ and CD8+ Glucose (mmol/l) 5.0 (4.5-5.4) 3.8 (2.9-4.4) 4.8 (4.5-5.2) counts or in the number of CD4+ and b-HB (mmol/l) < 0.1 2.6 (1.6-3.8)3 < 0.1 CD8+ lymphocytes as a percentage of Leptin (ng/ml) 12.5 (6.4-20) 6.6 (4.3-8.2)3 12.8 (8.2-20) total lymphocytes at any time point (Ta- IGF-1 (nmol/l) 16.6 (12-21.1) 9.0 (1.0-12.1)3 15.7 (11.9-22.2) ble III). 1Median (95% confidence intervals); n = 13. Significantly different from baseline (Wilcoxon signed rank) 2p < 0.02, 3p < 0.005. Mitogen induced activation of CD4+ and CD8+ cells To assess whether the ketogenic diet had Table III. Effects of ketogenic diet upon total lymphocyte, CD4+ and CD8+ counts and percentage of CD4+ and CD8+ cells expressing CD69 in mitogen stimulated whole blood.1 an inhibitory effect upon lymphocyte activation, we studied the expression of Baseline Day 7 of After 2 weeks the early activation marker, CD69, on ketogenic diet of re-feeding CD4+ and CD8+ lymphocytes after mi- togenic stimulation in whole blood. We Leukocyte count (109/l) NS 7.8 (5-10.4) 7.0 (5.2-12.2) 6.9 (5.5-11.3) used whole blood in preference to tradi- Lymphocyte count (109/l) 1.58 (1.4-2.03) 1.49 (1.10-1.97)2 1.54(1.24-2.52) tional methods involving the isolation of CD4+ (cells / ml) NS 690 (480-900) 720 (400-900) 700 (470-900) mononuclear cells as we considered that CD8+ (cells / ml) NS 250 (190-340) 270 (140-360) 250 (160-300) changes which might occur in the plasma CD4+ (% of total lymphocytes) NS 41 (30-56) 42 (31-51) 42 (26-53) concentrations of hormones and fuels in CD8+ (% of total lymphocytes) NS 15 (12-19) 15 (11-20) 15 (9-17) response to the ketogenic diet could have CD4+CD69+ (% of CD4+ cells) NS 72 (61-81) 69 (59-84) 74 (63-81) important effects upon T-cell function. CD8+CD69+ (% of CD8+ cells) NS 71 (58-86) 69 (61-82) 73 (64-83) However, despite the changes we observed in the concentrations of IGF-1, 1Median (95% confidence intervals); n = 13 2 leptin and cortisol, we could not detect Significantly different from baseline (Wilcoxon signed rank) p < 0.03. NS = no significant changes at any timepoint. any significant changes in CD69 expres-

212 Hypothalamus-pituitary-adrenocortical and -gonadal axis in RA / M. CutoloKetogenic diet and rheumatoid arthritis / D.A.EDITORIAL Fraser et al. ing upon disease activity without com- period. The significant weight loss which relevant immunosuppression did not oc- promising the patient’s nutritional status. we observed at day 7 of the ketogenic cur. However, as we did not observe any diet was most likely associated with a The findings of Lord et al. that leptin changes in either the laboratory or clini- temporary negative fluid balance sec- plays a pivotal role in fasting-induced cal variables, with the exception of morn- ondary to natriuresis. Sodium excretion immunosuppression were carried out in ing stiffness, this hypothesis does not during carbohydrate restriction has been mice, where a 2-day fast reduced body appear to hold. We must, however, also shown to mimic that which occurs in weight by as much as 20-30% (18). Their take into account that we compared the response to fasting (23). findings may thus be more relevant to effects of a ketogenic diet with results Although we found a significant increase chronic malnutrition in humans, where from our earlier investigation into the in serum cortisol concentrations after the a greater loss of body mass occurs. In effects of fasting, i.e. the two groups ketogenic diet, the increase was appar- our study, the ketogenic diet reduced were not directly compared. Although ently not sufficient to reduce disease leptin levels by 47% as compared to a there were no significant differences be- activity or influence the T-lymphocyte 67% reduction after sub-total fasting (4). tween the patients in the fasting study number or functioning. The 15% in- We therefore cannot rule out the possi- and the present study with regard to age crease in serum cortisol we observed is bility that the greater decrease in serum or disease duration, additional confound- comparable with the 22% increase ob- leptin after sub-total fasting is of rel- ing factors may have had an important served by Phinney et al. in healthy sub- evance to the observed immunosuppres- bearing on the outcomes of the respec- jects after a ketogenic diet (24). We did sion. tive regimes. We must also take into ac- not determine 24-hour cortisol concen- The 46% decrease in serum IGF-1 lev- count the fact that caution should be ob- trations, and therefore cannot estimate els observed at the conclusion of the ke- served when drawing negative conclu- the true magnitiude of the changes in togenic diet approached the 55% reduc- sions based on a small number of pa- cortisol production induced by the ke- tion we previously noted after 7 days of tients. togenic diet. However, our results do not acute starvation (4). Serum IGF-1 is a To achieve our objectives and to encour- support the idea that a counter-regula- marker of nutritional status and has been age patient compliance, it was necessary tory increase in serum cortisol concen- shown to correlate with nitrogen balance to design a diet which was palatable yet trations in response to carbohydrate re- (28). The fall we observed in IGF-1, de- fulfilled the required criteria regarding striction has immunomodulatory or clini- spite the provision of adequate dietary carbohydrate content. At the conclusion cally relevant effects in RA patients. protein and calories, may be due to the of the ketogenic diet, all patients stated Nevertheless, it is quite possible that fact that a ketogenic diet induces a nega- that the diet was palatable and that they other factors associated with the stress tive nitrogen balance for a period of were able to complete the protocol as of acute energy deprivation stimulate around one week before adaptation oc- instructed. The use of a ketogenic diet is additional cortisol secretion in fasting curs (24). well documented in the treatment of RA patients. The marked reduction we observed in childhood epilepsy and appears to have A fall in serum leptin has been impli- serum IGF-1 could have immunomodu- no apparent side effects when followed cated as a pivotal factor in starvation- latory effects, as IGF-1 is known to en- over a period of several years (21). The induced immunosuppression (18). In- hance T-cell proliferation (29). It has significant increases in b-HB and the creasing concentrations of leptin, levels previously been suggested that the T-cell decrease in glucose concentrations to of which are directly correlated with fat suppression observed during fasting may levels similar to those observed in re- mass (25), enhance T-cell proliferation be secondary to falls in serum IGF-1 (6). sponse to fasting levels provide some and promote a Th1 response in vitro (18). However, as we did not observe any indirect evidence that the patients did An increased Th1 response may be det- immunological effects of the ketogenic consume the provided diet as claimed rimental in RA (26). These findings are diet, we do not suspect the acute falls in (22). also of interest in relation to the finding IGF-1 of the magnitiude we observed As the patients were allowed to live at that a BMI > 30 kg/m2 is associated with induce decreases in T-cell numbers and home for the duration of the study, it is an adjusted odds ratio of 3.74 for devel- function. possible that the increase in b-HB levels oping RA (27). Our finding that the ke- In summary, using a ketogenic diet we resulted from the patients simply fast- togenic diet induced falls in leptin lev- have found metabolic and hormonal ad- ing. However, we observed that the me- els approaching those observed in re- aptations to carbohydrate restriction do dian body weight was only 1 kg less than sponse to sub-total fasting in the absence not induce clinical improvements or im- baseline at 2 weeks after the ketogenic of changes in T-cell number or activa- munomodulatory changes in RA pa- diet as compared to the 3.5 kg weight tion does not support the idea that leptin tients. The lack of any clinical or labo- loss which was still present 2 weeks af- plays a pivotal role in the fasting-induced ratory improvements in response to the ter sub-total fasting (4). Given that the suppression of CD4+ cell activation and ketogenic diet suggest that the reductions re-feeding diets were similar in both stu- numbers we previously observed (4). we observed in serum IGF-1 and leptin dies, it would appear that a much greater The unchanged laboratory and clinical do not appear to have clinically relevant calorie deficit occurred during the fast- variables in response to the ketogenic effects. Although we cannot yet identify ing period compared to the ketogenic diet diet would also suggest that clinically specifically which factors mediate either

213 Ketogenic diet and rheumatoid arthritis / D.A. Fraser et al. the immunosuppression or the clinical starvation. J Clin Endocrinol Metab 1997; 82: 18. LORD G, MATARESE G, HOWARD J, BAKER R, effects of fasting in RA patients, addi- 1177-80. BLOOM S, LECHLER R: Leptin modulates the 7. WOODWARD B: Protein, calories and immune T-cell immune response and reverses starva- tional factors associated with acute en- defences. Nutr Rev 1998; 56: S84-S92. tion-induced immunosuppression. Nature ergy and/or protein deficiency may be 8. KRAMER T, MOORE R, SHIPPEE R et al.: Ef- 1998; 394: 897-901. important. fects of food restriciton in military training on 19. ARNETT FA, EDWORTHY SM, BLOCH DA et T-lymphocyte responses. Int J Sports Med al.: The American Rheumatism Association Acknowledgements 1997; 18: S84-S90. 1987 revised criteria for the classification of 9. VAN DE LAAR MAFJ, VAN DER KORST JK: rheumatoid arthritis. Arthritis Rheum 1988; 31: We wish to thank Dr. Anne-Marit Food intolerance in rheumatoid arthritis. I. A 315-24. Selvaag at the Centre for Rheumatic Dis- double blind, controlled trial of the clinical 20. STEINBROCKER O, TRAEGER CH, BATTER- eases for her clinical assistance, and the effects of elimination of milk allergens and azo MAN RC: Therapeutic criteria in rheumatoid staff personnel at the Centre for Rheu- dyes. Ann Rheum Dis 1992; 51: 298-302. arthritis. JAMA 1949; 140: 659-62. 10. HAUGEN M, KJELDSEN-KRAGH J, FØRRE Ø: 21. KINSMAN SL, VINING EPG, QUASKEY SA, matic Diseases, the Department of Clini- A pilot study of the effect of an elemental diet MELLITS D, FREEMAN JM: Efficacy of the cal Chemistry and the Department of in the management of rheumatoid arthritis. Clin ketogenic diet for intractable seizure disorders: Endocrinology, The National Hospital Exp Rheumatol 1994; 12: 275-9. Review of 58 cases. Epilepsia 1992; 6: 1132-6. 11. KLEIN S, WOLFE R: Carbohydrate restriction 22. CARLSON MG, SNEAD WL, CAMPBELL PJ: for their technical assistance. regulates the adaptive response to fasting. Am Fuel and energy metabolism in fasting humans. J Physiol 1992; 262: E631-6. Am J Clin Nutr 1994; 60: 29-36. References 12. FERY F, BOURDOUX P, CHRISTOPHE J, BA- 23. BOULTER P, HOFFMAN R, ARKY R: Pattern of 1. SKÖLDSTAM L, LARSSON L, LINDSTRÖM FD: LASSSE E: Hormonal and metabolic changes sodium excretion accompanying starvation. Effects of fasting and lactovegetarian diet on induced by an isocaloric isoproteinic ketogenic Metab Clin Exp 1973; 22: 675-83. rheumatoid arthritis. Scand J Rheumatol 1979; diet in healthy subjects. Diabetes Metab 1982; 24. PHINNEY S, BISTRIAN B, WOLFE R, BLACK- 8: 249-55. 8: 299-305. BURN G: The human metabolic response to 2. HAFSTRÖM I, RINGERTZ B, GYLLENHAM- 13. DE FEO P, PERRIELLO G, TORLONE E et al.: chronic ketosis without caloric restriction: Phy- MER H, PALMBLAD J, HARMS-RINGDAHL M: Contribution of cortisol to glucose counter- sical and biochemical adaptation. Metabolism Effects of fasting in disease activity, neutrophil regulation in humans. Am J Physiol 1989; 257: 1983; 32: 757-68. function, fatty acid composition, and leuko- E35-E42. 25. CONSIDINE R: Weight regulation, leptin and triene biosynthesis in patients with rheumatoid 14. LANGFORT J, PILIS W, ZARZECZNY R, NAZAR growth hormone. Horm Res 1997; 48 (Suppl. arthritis. Arthritis Rheum 1988; 31: 585-92. K, KACIUBA-USCILKO H: Effect of low-car- 5): 116-21. 3. KJELDSEN-KRAGH J, HAUGEN M, BORCH- bohydrate-ketogenic diet on metabolic and 26. MIOSSEC P, CHOMARAT P, DECHANET J: By- GREVINK CF et al.: Controlled trial of fasting hormonal responses to graded exercise in men. passing the antigen to control rheumatoid ar- and one-year vegetarian diet in rheumatoid ar- J Physiol Pharmacol 1996; 47: 361-71. thritis. Immunol Today 1996; 17: 170-3. thritis. Lancet 1991; 338: 899-902. 15. BODEN G, CHEN X, MOZZOLI M, RYAN I: Ef- 27. SYMMONS D, BANKHEAD C, HARRISON B et 4. FRASER D, THOEN J, RESELAND J, FØRRE Ø, fect of fasting on serum leptin in normal hu- al.: Blood transfusion, smoking and obesity as KJELDSEN-KRAGH J: Decreased CD4+ lym- man subjects. J Clin Endocrinol Metab 1996; risk factors for the development of rheuma- phocyte activation and increased IL-4 produc- 81: 3419-23. toid arthritis: Results from a primary care- tion in peripheral blood of rheumatoid arthri- 16. GRINSPOON S, ASKARI H, LANDT M et al .: based incident case-control study in Norfolk, tis patients after acute starvation. Clin Rheum- Effects of fasting and glucose infusion on ba- England. Arthritis Rheum 1997; 40: 1955-61. atol 1999; 18: 394-401. sal and overnight leptin concentrations in nor- 28. ISLEY WL, UNDERWOOD LE, CLEMMONS 5. HOLM G, PALMBLAD J: Acute energy depri- mal-weight women. Am J Clin Nutr 1997; 66: DR: Dietary components that regulate serum vation in man: effect on cell-mediated immu- 1352-6. somatomedin-C concentrations in humans. J nological reactions. Clin Exp Immunol 1976; 17. TUOMINEN J, EBELING P, LAQUIER F, HEI- Clin Invest 1983; 71: 175-182. 25: 207-11. MAN M, STEPHENS T, KOIVISTO V: Serum 29. JOHNSON E, JONES L, KOZAC R: Expression 6. SAVENDAHL L, UNDERWOOD L: Decreased leptin concentration and fuel homeostasis in and function of insulin-like growth factor re- interleukin-2 production from cultured periph- healthy man. Eur J Clin Invest 1997; 3: 206- ceptors on anti-CD3-activated human T lym- eral blood mononuclear cells in human acute 11. phocytes. J Immunol 1992; 148: 63-71.

214