QJM vol. 97 no. 11 © Association of Physicians 2004; all rights reserved.
Unexplained swelling symptoms in women (idiopathic oedema) comprise one component of a common polysymptomatic syndrome
From the 1University Department of Human Nutrition, Glasgow Royal Infirmary, 2 Department of Psychiatry, Hairmyres Hospital, East Kilbride and 3 University Department of Public Health Medicine, Glasgow, UK
Received 16 February 2004 and in revised form 5 July 2004
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
Background: Non-menstrually-related swelling symptoms (idiopathic oedema) are common in women. The community prevalence of such symptoms, their association with other symptoms, and their underlying aetiology, are uncertain.
Aim: To determine the community prevalence of swelling symptoms and the independent contributions of major risk factors.
Design: Comparison of major risk factors in women with and without swelling symptoms.
Methods: We assessed 196 women attending a menopause clinic, 201 women attending a fracture clinic and 201 women attending their general practitioner. Each documented family histories of swelling symptoms and diabetes mellitus, age, height and current weight. Women attending the menopause and fracture clinics also completed Visual Analogue Symptom (VAS) scales documenting the perceived severity of swelling symptoms, and of 20 affective, somatic and functional autonomic symptoms. The independent contributions of risk factors to swelling symptom risk were estimated by logistic regression analysis.
Results: Of those attending a fracture clinic or their general practitioner, 28% and 33%, respectively, experienced non-menstrually-related swelling symptoms in the month before interview. Severe (RR 43, 95%CI 16–112, p < 0.001) and moderate (RR 7.8, 95%CI 4–15, p < 0.001) affective symptoms, a family history of swelling symptoms (RR 4.5, 95%CI 2.3–8.8, p < 0.001) and a body mass index (BMI) 
25 kg/m2 (RR 4.8, 95%CI 2.5–8.9, p < 0.001) were significantly associated with the presence of mild to severe swelling symptoms (VAS 1–9). The prevalence of swelling symptoms increased from 8% in women with no risk factors to 100% in women with three risk factors, which included severe affective symptoms.
Discussion: Affective symptom severity provides the principal independent contribution to swelling symptom risk. BMI 25 and a family history of swelling symptoms provide smaller independent contributions. The nature of the mechanisms underlying these associations remains uncertain.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Most general practitioners and hospital clinicians are familiar with women who complain of non-menstrually-related swelling of their face, hands, abdomen and feet. Physical examination usually shows no evidence of pitting oedema and investigation reveals no evidence of cardiac, hypoproteinaemic or obstructive causes for oedema. Swelling symptoms without obvious cause are usually called idiopathic oedema (and are also known as cyclical oedema, periodic oedema or the fluid retention syndrome).1–4 In the present paper, the simpler descriptive term ‘swelling symptoms’ is used.
Several aspects of the clinical features of swelling symptoms remain unresolved. With the exception of a small study by one of the authors,4 there are no documented surveys of the community prevalence of swelling symptoms. The severity of the discomfort experienced is disproportionate to objective evidence of swelling, although less obvious swelling of the face, fingers, ankles and abdomen is often evident to the complainant and close relatives familiar with her normal appearance, particularly in the evening.2–4 Median self-recorded daily weight gain over 4 weeks in idiopathic oedema patients with severe swelling symptoms was modest at 0.89 kg, indicating average daily fluid retention of <1 l.5 While maximum daily weight variation recorded by idiopathic oedema patients was significantly greater than in a control group of women,5,6 Thorn's widely accepted operational criterion for the diagnosis of this condition (diurnal weight variation exceeding 1.4 kg)1 provided poor between-group discrimination, with false negative and positive error rates of 59% and 18%, respectively, for cases and controls. The perception of swelling symptom severity appears to be independent of the magnitude of the fluid shift that has been its defining abnormality.
Subcutaneous swelling of the tissues of the face, hands, breasts, abdomen and feet appears to result from increased capillary permeability with secondary hypovolaemia compensated by oliguria, sodium and water retention, and weight gain aggravated by walking or standing.1,3,7,8 Patients with idiopathic oedema, in whom recent diuretic administration was excluded, demonstrated no consistent abnormalities of serum urea or electrolytes, renin, aldosterone, catecholamines, thyroid hormone or atrial natriuretic peptide concentrations.9–14 Swelling symptoms are common in post-menopausal women and have been described in prepubertal boys and girls,15 excluding an aetiological role for ovarian hormones. The hypothesis that idiopathic oedema usually results from diuretic or laxative use is unsupported by its occurrence in women who deny diuretic use, and by failure to detect diuretics and laxatives in unannounced urine tests in most of these patients.11,12 Despite their widespread use, a literature search revealed no published controlled trials which assessed the efficacy of diuretics in the treatment of swelling symptoms (or of any other drug treatment advocated for this common complaint).
In 1974, a 47-year-old woman was referred to one of the authors (MGD) with severe swelling symptoms. Investigation revealed that her four daughters, five grandchildren and nine other female relatives were similarly affected. Between 1975 and 2001, the clinical features of over 1100 women referred to a clinic for ‘idiopathic oedema’ were examined. Referrals came from primary care, and most clinical specialties in West of Scotland hospitals; the largest proportion came from menopause and gynaecological clinics, to which women with a range of affective and somatic symptoms were referred, frequently accompanied by swelling symptoms. This large database established that swelling symptoms formed one component of a syndrome comprising affective, somatic and functional autonomic symptoms. The latter included symptoms fulfilling diagnostic criteria for the irritable bowel syndrome, urge frequency of micturition and neurogenic syncope and presyncope.5 Weight gain, obesity and family histories of swelling and diabetes mellitus were significantly associated with swelling symptoms.2,4,15 A case-control study demonstrated that patients with idiopathic oedema exhibited significantly more evidence of affective dis-turbance than other female hospital outpatients.16
We investigated the community prevalence of swelling symptoms and quantified the statistical contributions of major risk factors to swelling symptom prevalence and severity in a risk-factor model for this common complaint.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Study groups
Since a true community prevalence study proved unfeasible, two groups of women attending their general practitioner or a follow-up fracture clinic provided surrogate estimates of the community prevalence of swelling symptoms. A group of women with complaints which included moderate to severe swelling symptoms were provided by the menopause clinic which had been the largest single source of referrals to the idiopathic oedema clinic noted above.
General practitioner group
In this initial prevalence study, 201 women were interviewed in order of presentation to their general practitioner with complaints common to this setting; none of those interviewed presented with a primary complaint of swelling symptoms.
Fracture clinic group
Estimates of swelling symptom prevalence in primary care settings may be inflated by the over-representation of women with associated affective symptoms. A further estimate of swelling symptom prevalence was therefore obtained from 201 women attending a follow-up fracture clinic as patients or accompanying patients. Since their attendance was determined by an unpredictable traumatic event involving a predominantly healthy population, this sample appeared likely to provide a reasonable estimate of swelling symptom prevalence in the normal female population.
Menopause clinic group
One hundred and ninety-six women were recruited from the menopause clinic. Clinic medical staff referred women in order of presentation to one of the authors (JBH) with non-menstrually-related swelling symptoms, usually accompanied by a range of other affective, somatic and autonomic symptoms. Women with swelling symptoms related to menstruation or hormone replacement therapy were excluded.
Investigations
General practitioner group
A structured interview by one of the authors (MGD) documented the presence or absence of non-menstrually-related swelling symptoms in the previous month, height, current weight and date of birth. Known family histories of swelling symptoms and diabetes mellitus and current and past diuretic use were also recorded.
Menopause and fracture clinic groups
Following an initial explanation by a research dietician (JBH), women in the menopause and orthopaedic groups completed a detailed questionnaire which documented their perceived severity of swelling, affective, somatic and functional autonomic symptoms. Most women found the questionnaire straightforward and usually completed it within 15–20 min. Self-rated swelling symptom severity unrelated to menstruation was recorded on ordinal Visual Analogue Symptom scales (VAS, 0–9) as all-site severity and as the severity of facial, hand, breast, abdominal and ankle swelling in the previous month. VAS scales of affective (poor concentration, irritability, anxiety, depression, insomnia, food cravings and panic attacks), somatic (fatigue, pain at all sites, breast pain, headache, thirst and visual blurring), and functional autonomic symptom severity (flushing, sweating, constipation, diarrhoea, urinary frequency, faintness and faints) were similarly recorded in random order to discourage repetitive recording of scores for related symptoms. Participants also completed Hospital Anxiety and Depression (HAD) scales and recorded known family histories of swelling and diabetes mellitus (first and all degrees of relationship), height, current weight, and date of birth. No age limit was imposed on selection, provided other diagnostic inclusion criteria were met.
In the fracture clinic and general practitioner groups, organic causes of swelling symptoms were excluded by a brief history, examination, and scrutiny of the outpatient or primary care record if required. Organic causes of swelling symptoms in the menopause clinic group were excluded by a clinical history and examination and appropriate investigations, as documented previously.4–6,11,15 All participants in the three groups gave informed consent to being interviewed and the Research and Ethical Committee of Lanarkshire Health Board approved the protocol for the study.
Statistical methods
The significance of univariate between-group differences was assessed by the Mann Whitney U test, that of proportional differences by the 
2 test, and that of univariate correlations between VAS scales by the non-parametric Spearman rank correlation coefficient. Assessment of the independent contributions of risk factors to swelling symptom severity in both groups was determined by logistic regression analysis. Risk factors comprised age, body mass index (BMI), Hospital Anxiety and Depression (HAD) scales, 20 VAS scales of affective, somatic and functional autonomic symptom severity, and family histories of swelling symptoms and diabetes mellitus. Models identified the best-fit combination of independent risk factors associated with all (VAS 1–9), mild (VAS 1–5), and moderate to severe (VAS 6–9) all-site swelling symptom severity, compared with zero swelling severity (VAS 0). The discriminating power of the logistic model was further examined by quantifying the proportions of swelling positive women (VAS 1–9) with no, single, and multiple risk factors.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Estimated community prevalence of swelling symptoms (Table 1)
In a previous study, 29 of 107 (27%) women interviewed by one of the authors (MGD) while attending their general practitioner gave a history of mild to severe non-menstrual swelling symptoms in the previous month. Thirteen women (12%) had consulted their general practitioner about swelling symptoms on a prior occasion.4
|
In this study, the prevalences of mild to severe swelling symptoms in the general practice and fracture clinic groups were 33% and 28%, respectively. The age and BMI distributions of the groups were similar, and there were no significant between-group differences in the prevalence of family histories of swelling symptoms and diabetes. In the general practice group, there were significantly greater prevalences of family histories of swelling symptoms and diabetes mellitus among swelling-positive than among swelling-negative women. In the fracture clinic group, only the prevalence of family histories of swelling symptoms was significantly greater in swelling-positive women.
In the general practice and fracture clinic groups, 31 (15%) and 14 (7%) women, respectively, had previously consulted their general practitioner about swelling symptoms; 15 (8%) and 7 (3%) women were currently taking a long-term diuretic.
Risk factors for swelling symptoms (Table 2)
Five women from the menopause group referred with possible swelling symptoms proved swelling-negative and were analysed with swelling-negative women from the fracture clinic group. The remaining swelling-positive women from the menopause group were significantly older, more overweight (BMI) and more anxious and depressed than swelling-negative women; they also had a significantly greater proportion of first degree and all relatives with swelling symptoms and diabetes mellitus than did swelling-negative women.
|
Swelling-positive women from the fracture clinic group, representing the community prevalence of swelling symptoms, were also significantly more overweight, anxious and depressed and had a significantly higher proportion of first degree and all relatives with swelling symptoms than swelling-negative women. They had significantly less severe swelling symptoms and were significantly lighter, younger and less anxious and depressed than swelling-positive women from the menopause clinic group.
In the total sample of 397 swelling-positive and -negative women, there were significant correlations between swelling symptom severity (VAS 0–9) and BMI (r = 0.366; p < 0.001), HAD Anxiety (r = 0.431; p < 0.001), HAD Depression (r = 0.528; p < 0.001), and age (r = 0.183; p < 0.001).
Affective, somatic and autonomic symptoms (Tables 3 and 4)
Mild to severe swelling symptoms were recorded by 97% of the menopause clinic group (five referred women proved swelling-negative) and by 28% of the fracture clinic group. The average prevalence of severe affective, somatic, and severe functional autonomic symptoms (VAS 1–9) was 81%, 80% and 50%, respectively, of the menopause clinic group vs. 44%, 48% and 17% of the fracture clinic group. The prevalence and severity of swelling symptoms, and of 19/20 affective, somatic and autonomic symptoms, were significantly greater in the menopause clinic group than in the fracture clinic group (Table 3).
|
|
When the groups were combined, there were significant univariate correlations between swelling symptom severity and the severity of all 20 affective, somatic and autonomic symptoms. Considered separately, there were significant correlations between swelling symptom severity and 18/20 symptoms in both the menopause and fracture clinic groups (Table 4).
Logistic regression analysis (Table 5)
VAS scores of perceived all-site swelling severity were superior to site-specific scores as the dependent variable in explanatory models of independent risk factors associated with swelling severity. Total affective, somatic and autonomic symptom scores were similarly superior to individual symptom scores. Stepwise logistic regression eliminated total somatic and autonomic symptom scores, a family history of diabetes, HAD scores and age from the final best-fit logistic models.
|
Three major independent risk factors were significantly associated with the presence and severity of swelling symptoms (VAS 1–9). In descending order of relative risk, these comprised the upper and middle tertiles of the total affective symptom score (respectively described as severe and moderate affective symptoms), a BMI
25 kg/m2, and a positive family history of swelling symptoms (Table 5). The relative risks of swelling symptoms associated with moderate and severe affective symptoms rose six-fold and 11-fold, respectively, from risks associated with mild swelling symptoms (VAS 1–5) to risks associated with moderate to severe swelling symptoms (VAS 6–9). Overweight relative risk rose by a third in women with moderate to severe swelling symptoms, compared to women with mild swelling symptoms; familial relative risk was unrelated to swelling symptom severity.
Contributions of risk factors to swelling symptom prevalence (Figure 1)
Eight per cent of women with no risk factors were swelling-positive (VAS 1–9). In women with a single risk factor, the proportion of women with swelling symptoms rose from 21% in women with an overweight risk factor (BMI 25) to 69% in women with severe affective symptoms. In groups with two risk factors, the proportion of women with swelling symptoms rose from 57% in women with both overweight and familial risk factors to 96% in women who were both overweight and had severe affective symptoms. In women with all three risk factors, swelling symptom prevalences in women with moderate or severe affective symptoms were, respectively, 96% and 100%.
|
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
‘Idiopathic oedema’ remains a rather mysterious and poorly understood entity. This may partly result from its obscure nomenclature (idiopathic, periodic, or cyclical oedema, or the fluid retention syndrome), and partly from the contrast between complaints of severe swelling and the absence of either objective evidence of oedema or of any consistent biochemical or endocrine diagnostic markers for the condition. The strategy of asking women to document the subjective severity of their swelling symptoms provided a fruitful alternative approach to the investigation of this common complaint.
Unlike most disease categories, which are distributed as dichotomous variables (e.g. myocardial infarction, thyrotoxicosis), swelling symptoms and the common affective, somatic, and autonomic symptoms which accompany them are distributed as continuous variables whose magnitude varies from minimal to moderate and severe. Swelling symptoms of all grades of severity at multiple sites, unrelated to menstruation, were recorded by 27%,4 33%, and 28% of three groups of women, reflecting the community prevalence of this symptom. The consistency of these findings suggests that slight to severe non-menstrually-related swelling symptoms are experienced by about 3 in 10 women. Identification of the factors which result in a proportion of women with swelling symptoms seeking medical attention for a condition that the majority accept as part of everyday experience is the subject of the present study.
The logistic regression model established that perceived swelling symptom severity was independently associated with the severity of seven co-existing affective symptoms. Smaller independent contributions to swelling symptom risk were provided by overweight and familial risk factors. The prevalence of swelling symptoms was strongly related to the presence of one or more of these risk factors. Statistical association does not prove causation; possible causal mechanisms associated with each risk factor are considered in turn.
The familial risk factor is consistent with the hypothesis that genetic factors are causally related to the expression of swelling symptoms. The role of inheritance was first clearly defined in a previous study of familial idiopathic oedema in prepubertal children.15 The mode of inheritance may be further clarified by an ongoing DNA linkage analysis of 72 male and female members of four kindreds with family histories of severe swelling, affective, somatic and autonomic symptoms over three generations.
Previous evidence suggests that the overweight risk factor is causally related to swelling symptom severity. Weight gain commonly precedes the onset of swelling symptoms, and weight reduction reduces both the severity of swelling symptoms and the amplitude of diurnal weight variation. Improvement is maintained while effective weight control continues, with the almost inevitable return of symptoms as weight is regained.2,4,5,17
A family history of diabetes was not independently associated with swelling symptoms in the logistic regression model. However, the significantly greater proportion of diabetic relatives among swelling-positive women from the menopause and general practice groups is consistent with previous reports of an increased prevalence of diabetic relatives in idiopathic oedema patients.2,4,5 Diabetic oedema at the onset of insulin treatment,18 in insulin-dependent patients with unstable metabolic control,19 and endogenous hyperinsulinaemia accompanying type 2 diabetes and obesity, all point to possible causal relationships between insulin and swelling symptoms. Detailed evidence of relationships between insulin and water and electrolyte metabolism have been reviewed previously.15
Possible mechanisms underlying the significant relationships between swelling and affective symptom severity are unclear. Women with zero or minimal swelling symptom severity (VAS 0–2) recorded similar median diurnal weight variation (0.86 kg daily) to that of women with maximum perceived severity (VAS 9; 0.89 kg daily).5 In contrast, as noted above, swelling symptom risk associated with severe affective symptoms was 11 times greater for moderate to severe swelling symptoms (VAS 6–9) than for mild swelling symptoms (VAS 1–5; Table 5). These findings imply that affective disturbance does not increase objective swelling severity, as documented by diurnal weight variation, but is strongly associated with amplification of perceived swelling severity. Whether this relationship is causal, or reflects the common amplification of swelling symptoms and significantly correlated affective, somatic and autonomic symptoms by unknown neuroendocrine transmitter/receptor mechanisms is unknown.
Evidence suggests that neuroendocrine function is abnormal in patients with moderate and severe swelling symptoms (idiopathic oedema). One study reported significantly different responses to hypothalamic releasing hormones (TSH- and LH-releasing hormones) in idiopathic oedema patients compared with controls.20 Two case-control studies reported abnormal antidiuretic hormone (vasopressin) responses to the administration of ethanol and a water load in the case group, suggesting abnormal neurohypophyseal function.8,21 Reports of severe oedema induced by metoclopramide22,23 and of uncontrolled trials documenting reduced diurnal weight variation following treatment with bromocriptine24,25 (a dopamine agonist) are consistent with abnormal central dopaminergic function in this syndrome. Several case reports have documented swelling symptoms and oedema induced by psychotropic drugs (phenothiazines,26–28 trazodone29 and lithium30) which modulate neuroreceptor/transmitter function. Clinical experience of the rapid induction of swelling symptoms within a few minutes of severe emotional disturbance and the significant relationship of exacerbations of swelling symptoms and increased diurnal weight variation to traumatic life events in patients with the swelling syndrome are consistent with the transmission of affective symptoms to efferent autonomic vasodilator pathways via limbic-hypothalamic pathways.4
As noted previously, indirect evidence suggests that increased capillary permeability causes the extravasation of water and electrolytes to extravascular tissues in patients with swelling symptoms. An arteriolar pre-capillary sphincter under autonomic control regulates blood supply to the capillary bed.3,8 Active vasodilator pathways under hypothalamic control that increase skin and muscle blood flow following emotional stress have been demonstrated experimentally.31 The ability of short-term emotional and physical stimuli to induce the rapid onset of swelling symptoms (emotional stress, high ambient temperatures, alcohol, exercise and prolonged standing), and the remission of symptoms when these precipitating factors are removed, indicates that this process is labile and reversible. The postulated role of the pre-capillary sphincter in modulating capillary permeability in women with swelling symptoms is reinforced by case reports of the induction of swelling by vasodilator drugs (isosorbide32 and nifedipine33) and its reduction by vasoconstrictor drugs (ephedrine and amphetamine3,8).
The affective, somatic and autonomic symptoms that are significantly related to swelling symptom severity are shared with overlapping symptom clusters identified by their presenting symptoms.34 They include fibromyalgia, chronic fatigue syndrome, the irritable bowel syndrome, urge frequency of micturition and premenstrual syndrome. Parallel analyses of these symptom clusters by consistent methodologies, such as those used in the present study, may facilitate a more inclusive taxonomy of these diagnostic entities. In providing a risk factor model for the swelling syndrome, the present study may also facilitate the recognition, investigation and management of a common group of poorly understood psychosomatic syndromes which constitute a significant part of the primary care and general hospital case-load.
|
Acknowledgments |
|---|
We are indebted to Dr David McKay Hart and his staff for their assistance in selecting patients attending the Stobhill Hospital Menopause Clinic, and to the Consultant Orthopaedic staff who kindly gave us permission to approach patients attending the Hairmyres Hospital follow-up fracture clinic. MGD is also indebted to Dr David Jamieson and his colleagues for permission to interview patients attending their Muirhead and Moodiesburn surgeries. We gratefully acknowledge the cooperation of the women who agreed to participate in the study.
|
Footnotes |
|---|
Address correspondence to Dr M.G. Dunnigan, 104 Beechwood Drive, Broomhill, Glasgow G11 7HH. e-mail: matthewdunnigan{at}aol.com
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
1. Thorn GW. Approach to the patient with ‘idiopathic edema’ or ‘periodic swelling’. JAMA 1968; 206:333–8.
2. Dunnigan MG. Fluid retention syndrome. Management in women. Update 1989; 15 June:1431–5.
3. Streeten DHP. Idiopathic edema: pathogenesis, clinical features and treatment. Metabolism 1978; 27:353–83.
4. Dunnigan MG. Recognition and management of the fluid retention (idiopathic or cyclical oedema) and premenstrual syndromes. In: McNaughton MC, ed. Medical Gynaecology. Edinburgh, Blackwell Scientific, 1985:27–55.
5. Dunnigan MG, Czapla K, Henderson JB, Pelosi AJ. The fluid retention syndrome: recognition, risk factors and management. J Br Menopause Soc 1999; 5:59–65.
6. Denning DW, Dunnigan MG, Tillman J, Davis JA, Forrest CA. The relationship between ‘normal’ fluid retention in women and idiopathic oedema. Postgrad Med J 1990; 55:363–6.
7. Edwards OM, Bayliss RIS. Idiopathic oedema of women. A clinical and investigative study. Q J Med 1976; 45:125–44.
8. Streeten DHP. Orthostatic disorders of the circulation. Mechanisms, manifestations and treatment. New York, Plenum Publishing Corporation, 1987.
9. Oelkers W, Rawer C, Wiederholt R, Schoneshofer M, Pallicki H. Katamnesen und endokrinologische Befunde bei Frauen mit idiopathischem Odem. Klin Wschr 1977; 55:495–502.
10. Streeten DHP. Idiopathic oedema. Lancet 1979; i:776.
11. Dunnigan MG, Denning DW, Henry JA, de Wolff FA. Idiopathic oedema and diuretics. Postgrad Med J 1987; 63:25–6.
12. Pelosi AJ, Czapla K, Duncan A, Henderson JB, Dunnigan MG. The role of diuretics in the aetiology of idiopathic oedema. Q J Med 1995; 88:49–54.
13. Czapla K, Farish E, Haynes ECM, Henderson JB, Pelosi AJ, Dunnigan MG. Thyroid function in idiopathic oedema. Clin Endocrinol 1996; 45:467–70.
14. Anderson GH Jr, Streeten DHP. Effect of posture on plasma atrial natriuretic hormone and renal function during salt loading in patients with and without postural (idiopathic) edema. J Clin Endocrinol Metab 1990; 71:243–6.
15. Dunnigan MG, Pelosi AJ. Familial idiopathic oedema in prepubertal children: a new syndrome. Q J Med 1993; 86:301–13.
16. Pelosi AJ, Sykes RA, Lough JRM, Muir WJ, Dunnigan MG. A psychiatric study of idiopathic oedema. Lancet 1986; ii:999–1002.
17. Dunnigan MG, Lawrence JR. Idiopathic oedema. Lancet 1979; i:776–7.
18. Leifer A. A case of insulin edema. JAMA 1928; 90:610–11.
19. Dunnigan MG. Unusual clinical manifestations of diabetes mellitus. Practitioner 1979; 222:321–30.
20. Young JB, Brownjohn AM, Chapman C, Lee MR. Evidence for a hypothalamic disturbance in cyclical oedema. Br Med J 1983; 286:1691–3.
21. Thibonnier M, Marchetti J, Corvol P, Menard J, Milliez P. Abnormal regulation of antidiuretic hormone in idiopathic oedema. Am J Med 1979; 67:67–73.
22. Ahmed S. Metoclopramide induced parkinsonism and severe edema. South Med J 1977; 77:1345.
23. Zumott B. Metoclopramide oedema. Ann Intern Med 1983; 98:557.
24. Zimlichman RR, Porath A, Sobel R. Bromocriptine-responsive idiopathic edema. Isr J Med Sci 1984; 20:264–6.
25. Dent RG, Edwards OM. Idiopathic oedema; a study of the effects of bromocriptine. Clin Endocrinol 1979; 13:75–80.
26. Vitz L, Shapiro MS, Shenkman S. Chlorpromazine induced fluid retention masquerading as idiopathic oedema. Br Med J 1987; 294:807–8.
27. Dunleavy DL. Phenelzine and oedema. Br Med J 1977; 1:1353.
28. Margolis J. Massive edema induced by thioridazine (Melleril), an unusual complication. J Am Geriat Soc 1972; 20:593–4.
29. Barnett J, Frances A, Kocsis J, Brown R, Mann JJ. Peripheral oedema associated with trazodone: a report of ten cases. J Clin Psychopharmacol 1985; 5:161–4.
30. Stancer HC, Kiwi R. Lithium carbonate and oedema. Lancet 1971; ii:985.
31. Greenfield ADM. Survey of the evidence for active neurogenic vasodilatation in man. Fed Proc 1966; 25:1607–10.
32. Rodger JC. Peripheral oedema in patients treated with isosorbide dinitrate. Br Med J 1981; 283:1365–6.
33. Opic LH. Fluid retention with nifedipine in antihypertensive therapy. Lancet 1986; ii:1456.
34. Wessley S, Nimnuan C, Sharpe M. Functional somatic syndromes: one or many. Lancet 1999; 354:936–9.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||