Alcohol, Estrogens, and Breast Cancer

Barnett Zumoff

To the editor:

In a paper published in JCEM, Straub et al. (1) reported on the interrelations in diabetics between atrial natriuretic factor (ANF), endogenous digoxin-like immunoreactive factors (DLIF), and autonomic abnormalities.

The authors state that "a group of natriuretic hormones such as ouabain or DLIF play an important role in the hormonal feedback system to control volume expansion" and that "DLIF is secreted from the adrenal glands and hypothalamus," and they conclude that their results "demonstrate close interactions between the autonomic nervous system and the two natriuretic hormones" (DLIF and ANF) and that in diabetics with autonomic dysfunction "higher levels of ANF may possibly compensate for the lack of the natriuretic DLIF" (1). We think that the paper conveys an unjustified impression of established knowledge in an area that still requires great caution, and we feel compelled to comment and amend a few imprecisions in the article.

DLIF and ouabain

The existence in mammals of endogenous counterparts to digitalis (we propose the less committing term: endogenous digitalis-like factors or EDLF) has been long postulated (see 2), but most of the fundamental issues (structure, sites and enzymatic pathways of biosynthesis, physiological or pathological roles) remain controversial or unresolved.

A substance indistinguishable by mass spectrometry from ouabain was reported in human plasma by Hamlyn and coworkers (3) in 1991, but the question of whether ouabain is circulating in blood is still under debate (2, 4–7). Recently, it has been demonstrated that the EDLF isolated by Hamlyn et al. (3) is identical to the factor isolated from bovine hypothalamus (8) and is probably an isomer of ouabain. Also, other substances have been identified as EDLF by mass spectrometry: digoxin from human urine (9) and 19-norbufalin from human lens (10).

Many investigators have used DLIF for EDLF (see 11). However, Goto et al. (11), in a review of this topic, admit that this approach has several problems and conclude that "RIA for digoxin per se is an unreliable estimate of digitalis-like activity and has misled the search for true EDLF". Strangely, Goto is the author to whom Straub (1) refers for DLIF as the natriuretic hormone involved in the control of volume expansion [see Ref 21 in (1)]. Naomi et al. (12) showed that an "immunochemical fingerprint" with several antidigoxin antisera was helpful in identifying a digoxin-like factor in humans, but unfortunately, they did not analyze the digoxin immunossay employed by Straub (1).

Hypothalamus and adrenals as sites of biosynthesis of EDLF

Our state of knowledge is insufficient to state that "DLIF is secreted from the adrenal glands and hypothalamus" (1). Yamada et al. (13) reported on a DLIF present in the hypothalamus, but others (14, 15) were unable to demonstrate cross-reactivities with antidigoxin antibodies in mammalian hypothalamus.

Measurable amounts of ouabain-like immunoreactivity have been reported in rat adrenal extracts and in cultures of bovine adrenocortical cells (16) and, in the awake dog, an increase of ouabain immunoreactivity has been reported in the adrenal vein (17). In contrast, Doris et al. (4) from studies with bovine and murine adrenocortical cells concluded as unlikely that adrenals produced ouabain in meaningful amounts. Furthermore, it has been reported (18) that there was no difference in plasma ouabain immunoreactivity in two patients with Cushing’s disease after bilateral adrenalectomy and no step-up in the adrenal vein compared with the inferior cava. Finally, plasma EDLF was unaffected by adrenal removal in normal and salt-loaded rats (19, 20), and adrenalectomy did not prevent EDLF increase caused by high sodium intake in rats (20). We also found that, in patients without adrenals, EDLF concentration in plasma as assessed by a biological assay (21) and by Du-Pont-NEN ouabain immunoassay (22), was similar to that of normotensive subjects.

DLIF as a natriuretic hormone

We are aware that the idea of endogenous digitalis has originated in renal physiology, but we feel that, at the present stage of knowledge, the role of EDLF in sodium balance is plausible but still unproven, and other functions may equally likely be affected by EDLF, from cataract formation (10) to vasoconstriction (23), to adaptation of postnatal life (24, 25).

DLIF assay

It has been demonstrated that the degree of interference by DLIF in the assay employed by Straub et al.(1), i.e. the Abbott TDX fluorescence polarization immunoassay (FPIA), is generally less than in other digoxin RIAs (26). Also, it differs from two other automatic digoxin immunoassays (Baxter-Dade Stratus II and Ciba Corning ACS) (27). Moreover, it has been repored that the treatment of serum recommended in the FPIA before the assay produced deglycosylated congeners of both digoxin and DLIF and that this breakdown may explain discrepancies with other studies (28).

Taken together, we feel that the study of Straub et al. (1) would have greatly benefited by a much more cautious, detailed, and rigorous introduction and discussion. We fear that studies like the one reported by Straub et al. (1) might generate in the non well-informed reader unjustified or even misleading expectations, which when contradicted by other (often similar) studies run the risk of discrediting this still unsettled but fascinating area of endogenous digitalis.

To the editor:

As Paci and Ghione said in their letter (above), many investigators have used immunological digoxin methods for the measurement of DLIF. For clinical use as in this study, specific RIAs are widely accepted to measure DLIF. The group of Balzan et al. (1) compared digoxin RIA with 3H-oubain displacement on erythrocyte membranes and found a good correlation of about r = 0.7, P < 0.0001 between both variables. In this article and in a later article by the same group (2), they "... confirmed the idea that radioimmunological methods may be useful in studying endogenous inhibitors of the sodium pump" (1). Indeed, the studies of this group (1, 2) convinced us to use an immunological assay for the measurement of DLIF in a clinical study set-up. Now, some years later they (Paci and Ghione) criticize their own recommendation to use immunological methods for measurement of DLIF.

In our work on atrial natriuretic and immunoreactive factors in diabetics (3), we referred to Goto et al. (4) but not to the paper mentioned by Paci and Ghione (reference 11 of their letter above). The Goto paper (4) discusses the natriuretic effect, and indeed, there is some controversy about the physiological action of oubain-like compounds. Nevertheless, several other investigators discussed DLIF or related substances known to be natriuretic. Furthermore, in paragraph 3 of the letter of Paci and Ghione they state "... the role of EDLF in sodium balance is plausible ..." Indeed, because of effects of the Na/K-pump, natriuretic effects are very plausible.

From the references cited by Paci and Ghione in paragraph 2, we still believe that DLIF or related compounds are produced in the adrenal gland and the hypothalamus (see Paci and Ghione’s references 13, 16, 17) (see references 11, 12 in our paper). All together there are five laboratories that find DLIF or related compounds in the hypothalamus or adrenal glands. Is this not a clear sign?

It may be that, with the use of TDX, sulfosalicylic acid alters the structure of digoxin and DLIFs. However, this may at best lead to false low levels of DLIFs and not to false high levels. It may be that DLIF serum concentration in our DLIF-positive patients would have been higher with an other determination technique.

To the editor:

A 42-yr-old acromegalic patient who developed acute pancreatitis following the exposure on separate occasions, to both lanreotide and octreotide, has been recently reported by Soule et al. (1). The authors speculated that long-acting somatostatin analogs might induce a smooth muscle spasm of the sphincter of Oddi with consequent short-lived pancreatic duct obstruction. On the other hand, Giusti et al. (2) in the authors’ response to Dr. Soule et al., referring to an extensive multicenter study (3), reported in 4 out of 27 patients with normal baseline amylase values, a slight increase of serum amylase levels with normal urinary amylase excretion during the first 6 months of lanreotide therapy, but no increased risk on 5 patients with basal serum amylase levels slightly above the upper normal limit. Moreover, several reports concern the relative effectiveness of somatostatin and octreotide in preventing risk of complications following pancreatic surgery for chronic pancreatitis (4) or in the treatment of acute pancreatitis (5). In fact, it has recently become clear that somatostatin, but not octreotide, is a useful and effective therapy for acute pancreatitis and for preventing complications following endoscopic retrograde cholangiopancreatography (6), because of differential effects on sphincter of Oddi motility (the native hormone relaxing and the analogue increasing its contractility).

We report a second acromegalic patient who developed acute pancreatitis in two separate exposures to both octreotide and lanreotide.

An acromegalic 29-yr-old man showed transient normalization of serum GH and insulin-like growth factor I levels following transsphenoidal approach in 1992. The patient was treated with octreotide (100 µg twice daily sc) since January 1994. No ultrasonographic or biochemical evidence of pancreatic disease was evident before octreotide treatment. Serum GH and insulin-like growth factor I values reversed to normal during the first day; the treatment was well tolerated until the third day, when abdominal pain, nausea, and vomiting developed and serum pancreatic amylase and lipase values raised to 1223 U/L (normal values 17–115) and 2107 U/L (normal values <190), respectively. Despite the continuation of octreotide administration, computed tomography (CT) scan and abdominal ultrasonography did not demonstrate morpho-volumetric and structural alteration of the pancreas or of other abdominal organs. Normal Vater’s ampulla with regular bile efflux were found at gastroduodenoscopy. Serum amylase and lipase levels reversed to normal in 3 weeks. Dietetic disorders or other inducing factors had been excluded.

After 13 months of treatment the patient complained again of abdominal pain, followed by serum pancreatic amylase and lipase levels increasing to 482 U/L and 622 U/L, respectively, without ultrasonographic evidence of gallbladder and pancreatic changes. The octreotide treatment was stopped and gamma knife radio-surgery was performed. Because of a persistent increase of serum IGF-1 levels, the patient was started on low-release somatostatin analog lanreotide treatment (30 mg im) in November 1996. Nine hours after the first injection however, abdominal pain and vomiting developed, associated with a sharp increase of serum amylase and lipase levels to 1065 U/L and 1885 U/L, respectively. No pancreatic lesion was apparent at abdominal CT scan. Serum pancreatic amylase and lipase values were still elevated (305 U/L and 595 U/L) 2 weeks later.

None of the other 14 acromegalic patients treated with lanreotide for almost 3 months has shown serum amylase levels above the upper normal limit, although in one case, a slight increase of pancreatic amylase values, but not of serum lipase levels, had been found during previous octreotide therapy.

We conclude that, although several adverse gastrointestinal events but not pancreatitis are currently reported during octreotide or lanreotide therapy, a slight risk of acute pancreatitis following somatostatin analogue treatment for acromegaly could exist, and we are in agreement with the other authors (1, 2), on the need for a careful pancreatic and gallbladder survey before and during the therapy.

To the editor:

To our surprise, Hunag et al. (1) related that half of their patients with primary aldosteronism initially sought medical attention because of muscle paralysis. The term paralysis, when applied to voluntary muscles, means loss of contraction. However, in medical parlance, paralysis sometimes means either partial or complete loss of strength. It is preferable to use paresis for slight and paralysis or plegia for severe loss of strength. Muscle strengths are scored as grade 0–5, according to the Medical Research Council Scale of Great Britain (2). Grade 2 indicates active movement with gravity eliminated; grade 3 indicates active movement against gravity. Under grade 3, patients cannot ambulate, causing them to seek medical help immediately. Huang et al. reported that 49% patients with primary aldosteronism had paralytic myopathy (he used this term). They did not mention how many patients had true paralysis, although an unexplained higher incidence of periodic paralysis had been observed in Chinese patients (3, 4). We question the assumption that mild muscle weakness would bring the patients to the hospital.

From 1985 to 1996, we had made diagnosis of primary aldosteronism in 79 patients (aged 23–73 yr; female 59, male 20) at National Taiwan University Hospital. Sixty cases were surgically proven to be unilateral aldosterone producing adenoma, and 19 cases were idiopathic hyperplasia. All patients had hypertension. Headache (53.3%) was the first leading symptom. One-third of patients denied any discomfort but suffered hypertension, nevertheless. Muscle weakness could be recalled in 35 cases (43%). In these patients, the weakness made difficult such activities as walking far, going upstairs, getting up from bed, or squatting. The symptoms were usually self-limited however, and most of the patients did not seek medical help for such symptoms. In contrast to Huang’s report, only 6 patients had true paralysis with profound hypokalemia (serum potassium, 1.9 \ 0.2 mEq/L), which brought them to the hospital. Primary aldosteronism was therefore suspected. Rhabdomyolysis without significant impairment of renal function was documented in 5 patients. There were also 5 patients with primary aldosteronism who first sought help because of stroke, and 1 patient whose hypokalemia was found during hospitalization for preeclampsia. In the remainders, primary aldosteronism was suspected because of hypertension and hypokalemia in 56 patients, abnormally high aldosterone/renin ratio with normokalemia in 10 patients, and adrenal lesion on CT scan in 1 patient. The serum potassium level of the patients with muscle weakness (including paralysis) was not different from those had no such symptom (2.41 \ 0.6 vs. 2.73 \ 0.58 mEq/L, respectively, P = 0.71).

Compared with the reports of others (3–5), the incidence of neuromuscular symptoms was not greater in the Taiwanese studies (49% in Huang’s and 43% in ours). Although neuromuscular symptoms because of hypokalemia is a common presentation in patient with primary aldosteronism (3–5), the clues leading to suspicious diagnosis are hypokalemia and/or high aldosterone/renin ratio in screening tests (6). A test of serum potassium is the simplest way to screen the disease, although normokalemia has been reported in various proportions. Omission of such testing in the initial evaluation of hypertensive patients may lead to misdiagnosis and late complication of this disease. In our series, hypertension with hypokalemia is the most common reason making the suspected diagnosis initially. The discrepancy of the initial symptoms leading to the diagnosis of primary aldosteronism between ours and Huang’s patients is not clear. However, it may be the result of different strategies of antihypertensive drugs or diet that may induce profound hypokalemia and paralysis.

Acknowledgments.

We thank the Lee Lin Hsiu Chin Kidney Research Fund for grant support of this manuscript.

To the editor:

In my recent editorial on this subject (1), in JCEM June, I made a very brief and quite peripheral comment that Dr. Susan Love, in her Op-Ed article in The New York Times (2), had misquoted the statistics about the extent of the increase in risk of breast cancer with estrogen administration in the Nurses Health Study report by Colditz, et al. (3) in 1990. It has been brought to my attention that Love was quoting not the 1990 paper, but a 1995 update report by the same research group (4). The update does indeed contain the figures she quoted and therefore disagrees somewhat with the earlier report, which did not contain those figures (note, however, that the 1995 report does not withdraw or modify in any way the 1990 report’s conclusion that alcohol consumption is required for there to be any increased risk of breast cancer with estrogen administration). I apologize for misunderstanding which of the Nurses Health Study reports Love was quoting and for therefore ascribing to her an error that she did not make. That, however, is irrelevant to and does not detract from the validity of the principal point of my editorial, namely, that the literature to date suggests strongly that postmenopausal women who consume less than 5 gm alcohol daily, or none at all, appear to be at no increase in risk of breast cancer when they take estrogen replacement therapy.

Footnotes

a Received March 31, 1997. Address correspondence to: Dr. Anna Paci, CNR Institute of Clinical Physiology, Via Savi 8, 56100 Pisa, Italy.

b Received April 30, 1997. Address correspondence to: Rainer H. Straub, M.D., Klinik und Poliklinik Für Innere Medizin I, Franz-Josef-Strauß-Allee 1, Regensburg, Germany D-93042.

c Received March 31, 1997. Address correspondence to: Salvatore Cannavò, M.D., Viale San Martino 380, 98122 Messina, Italy.

d Received April 21, 1997. Address correspondence to: Bor-Shen Hsieh, M.D., No. 7, Chung-Shan South Road, Department of Medicine, National Taiwan University Hospital, Taipei, Taiwan.

References

  1. Zumoff B. 1997 Editorial: The critical role of alcohol consumption in determining the risk of breast cancer with postmenopausal estrogen administration. J Clin Endocrinol Metab. 82:1656–1658.[Free Full Text]
  2. Love S. 1997 Sometimes Mother Nature Knows Best. New York Times. Op-Ed page, Thursday, March 20.
  3. Colditz GA, Stampfer MJ, Willett WC, Hennekens CH, Rosner B, Speizer EE. 1990 Prospective study of estrogen replacement therapy and risk of breast cancer in postmenopausaul women. JAMA 264:2648–2653.
  4. Colditz GA, Hankinson SE, Hunter DJ, et al. 1995 The use of estrogens and progestins and the risk of breast cancer in postmenopausal women. N Engl J Med. 332:1589–1593. [Abstract/Free Full Text]




This Article
Full Text (PDF)
Submit a related Letter to the Editor
Purchase Article
View Shopping Cart
Alert me when this article is cited
Alert me when eLetters are posted
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Request Copyright Permission
Google Scholar
Articles by Zumoff, B.
Articles citing this Article
PubMed
PubMed Citation
Articles by Zumoff, B.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Endocrinology Endocrine Reviews J. Clin. End. & Metab.
Molecular Endocrinology Recent Prog. Horm. Res. All Endocrine Journals