Effects of a Magnesium-Free Dialysate on Magnesium Metabolism During Continuous Ambulatory Peritoneal Dialysis

ORIGINAL INVESTIGATIONS

Effects of a Magnesium-Free Dialysate on Magnesium Metabolism During Continuous Ambulatory Peritoneal Dialysis Gaurang M. Shah, MD, Robert L. Winer, MD, Ralph E. Cutler, MD, Allen I. Arieff, MD, William G. Goodman, MD, John W. Lacher, MD, Patricia Y. Schoenfeld, MD, Jack W. Coburn, MD, and Arthur M. Horowitz, PhD • While the use of magnesium-containing compounds is usually contraindicated in dialysis patients, the risk of toxicity from hypermagnesemia can be reduced by lowering the magnesium concentration in dialysate. We examined the effects of a magnesium-free dialysate on both serum magnesium level and the peritoneal removal rate of magnesium over 12 weeks. in 25 stable patients undergoing continuous ambulatory peritoneal dialysis (CAPO). After 2 weeks, the serum magnesium level decreased from 2.2 to 1.9 mg/dL (0.9 to 0.8 mmol/L) (P < .02) and the peritoneal removal rate Increased from 66 to 83 mg/d (2.8 to 3.5 mmol/d) (P < .05), with both values remaining stable thereafter. There was a strong association between these parameters (r = - 0.62, P < .05), suggesting that the serum magnesium level decreased as a result of the Initial Increased peritoneal removal rate. For an additional 4-week period, a subgroup of nine patients received magnesium-containing, phosphate binding agents instead of those containing only aluminum. During this phase, serum Inorganic phosphorus was well controlled. The serum magnesium level increased only from 1.8 to 2.5 mg/dL (0.7 to 1.0 mmollL) (P < .05), due in great part to the concomitant 41% rise in peritoneal magnesium removal from 91 to 128 mg/d (3.8 to 5.3 mmol/d) (P < .05). No toxicity was noted during the entire 16-week study period, nor did serum calcium change. Thus, serum magnesium levels remained within an acceptable range as magnesium-containing phosphate binders were given through the use of magnesium-free peritoneal dialysate. Such a regimen has intriguing possibilities for the reduction of the oral aluminum burden to dialysis patients, with obvious Implications for the aluminum-related disorders observed in many of these patients. © 1987 by the National Kidney Foundation, Inc. INDEX WORDS: Magnesium; peritoneal flux; continuous ambulatory peritoneal dialysis.

D

ISORDERS of divalent cation metabolism are quite common in patients with end-stage renal disease (ESRD). In addition to the wellknown changes in calcium metabolism, there is impaired magnesium excretion and mild elevations of serum magnesium in the majority of patients. I ESRD patients have increased total body magnesium, primarily in bone,2 and this abnormality has been postulated to playa role in the genesis of renal osteodystrophy.2 Extraosseous complications of uremia may also be related to excess magnesium, since this cation has been shown to be an integral component of the microcrystalline compound in visceral calcifications. 3 Obviously, the administration of magnesium-containing comFrom the l-eterans Administration Medical Centers at Long Beach. Loma Linda. San Francisco. Sepulveda. and Los Angeles. CA; San Francisco General Hospital. CA; Presbyterian Hospital. Denver. CO; Abbott Laboratories. North Chicago. IL; the University of California at Irvine. Los Angeles. and San Francisco; Loma Linda University; and University of Colorado. Denver. Address reprint requests to Gaurang M. Shah. MD. Nephrology Section (III N) . 1/,4 Medical Center. 5901 E 7th Street. Long Beach. CA 90822. © 1987 by the National Kidney Foundation. Inc. 0272-638618711004-0002$3.0010 268

pounds may increase the serum and tissue levels of magnesium, with the potential of serious toxicity. 4 In hemodialysis patients , serum and tissue magnesium levels are related in part to the dialysate magnesium concentration. It has been shown that the magnesium level in serum,5 bone,6-8 and skin9 may be reduced by decreasing the magnesium concentration in dialysate from a level of 1.0 to 1.5 mEq/L (0.5 to 0.75 mmol/L) to 0.5 mEq/L (0.25 mmollL) or less. This modification of the dialysate may cause an improvement in bone mineralization,6 nerve conduction velocity,1O and pruritus . II Altered phosphorus metabolism and hyperphosphatemia are common in ESRD patients. 12 The latter is regularly managed with oral phosphate binding agents, most of which contain aluminum. However, some aluminum is absorbed from the gut,13 and this aluminum is deposited in various tissues in patients with impaired renal function. Recent studies indicate that excessive aluminum accumulation can be toxic, leading to the syndromes of dialysis encephalopathy\3 and osteomalacia. 14 Since water purification can virtually eliminate aluminum from dialysate, the major source of aluminum accumulation in ESRD pa-

American Journal of Kidney Diseases. Vol X. No 4 (October). 1987: pp 268-275

269

MAGNESIUM-FREE DIALYSIS IN CAPO PATIENTS

tients is from the aluminum-containing phosphate binders.13 Therapeutic alternatives to reduce the oral load of aluminum in ESRD patients include the use of phosphate binders that contain calcium, magnesium, or a combination of smaller amounts of aluminum and magnesium. 13,15,16 However, the long-term efficacy and safety of these substances, particularly with respect to hypercalcemia and hypermagnesemia, have not been determined. While the use of aluminum-magnesium binders may present a lower load of both ions, there may be a risk of hypermagnesemia. If the dialysate magnesium concentration is reduced, then the use of magnesium-containing phosphate binders might be a reasonable alternative to compounds that contain only aluminum. Magnesium metabolism has not been extensively investigated in patients undergoing continuous ambulatory peritoneal dialysis (CAPD). Early studies showed that mild hypermagnesemia, with serum magnesium levels in the range of 3.0 mgl dL (1.2 mmollL), occurred with the use of a dialysate magnesium concentration of 1.5 mEq/L (0.75 mmollL)Y,18 By lowering the dialysate magnesium to 0.5 mEq/L (0.25 mmollL), Nolph et al reduced the serum magnesium level to normal,19 and this dialysate concentration is now accepted as the standard for peritoneal dialysate. The present study was designed to examine the safety and efficacy of a magnesium-free dialysate in CAPD patients. We specifically assessed the effects of this dialysate on serum magnesium concentration and peritoneal magnesium flux, as well as the clinical manifestations of alterations in magnesium metabolism. In addition, we observed the clinical and metabolic effects of the short-term administration of aluminum-magnesium containing phosphate binders in a subset of the patients managed with magnesium-free dialysate. METHODS This multi-center study involved 25 stable patients who had undergone CAPO for at least 2 months using a dialysate magnesium concentration of 0.5 mEq/L (0.25 mmoIlL), There were 20 men and five women with a mean age of 53 years, Clinical diagnoses included chronic glomerulonephritis (12), nephrosclerosis (4), diabetic nephropathy (2), polycystic kidney disease (3), and miscellaneous disorders (4), The subjects were enrolled from the outpatient clinics of the Veterans Administration Medical Centers in Long Beach, Lorna Linda, San Francisco, and Sepulveda, CA; San Francisco General Hospital; and Presbyterian Hospital, Denver, CO, The research protocol was approved by the Institutional Review Boards at each

facility, and informed consent was obtained from all patients before participation in the study, The study was divided into three phases, Phase I consisted of a 4-week control period with use of a dialysate magnesium concentration of 0,5 mEq/L (Inpersol LM, Abbott Laboratories, North Chicago, IL; or Dianeal II, Travenol-Baxter Laboratories, Deerfield, IL), All patients who completed phase I were entered in phase II, an experimental period of 12 weeks duration, with CAPO carried out using magnesium-free dialysate (Inpersol ZM, Abbott Laboratories, North Chicago, IL), The composition of the two dialysates is shown in Table I, Patients were withdrawn from the study if serum magnesium fell below 1.3 mg/dL. Six patients developed peritonitis and they were treated according to the protocol of the individual institution, Blood and dialysate sampling were temporarily suspended, but were resumed after the peritonitis cleared, Nine patients entered phase III, and received a magnesium supplement for 4 weeks, The magnesium-free dialysate was continued and the oral phosphate binders containing only aluminum were replaced by ones that contained both aluminum and magnesium, Six patients recei~ed magaldrate (Riopan, Ayerst Laboratories, New York) in an average dose of2.4 g/d (range 1.2 to 3,6 g/d), replacing the 3.0 g/d of aluminum hydroxide during phase II. Three patients were given a combination of aluminum and magnesium hydroxides (Mylanta II, Stuart Pharmaceuticals, Wilmington, DE) in a dose of 4.8 g/d (range 2.4 to 7,2 g/d) which replaced 6,0 g/d of aluminum hydroxide ingested during phase II. Patient compliance was assessed by historical determination of phosphate binder ingestion and there were no apparent differences between phases II and III, These substitutions reduced the elemental aluminum load by 75 % in the group receiving magaldrate and 22 % in the group on aluminum-magnesium hydroxide, and increased the mean elemental load of magnesium by 0,7 and 2.0 g/d in the two groups, respectively, This protocol was not designed as a balance study, and neither dietary intake nor the nondialytic losses of divalent cations were measured, Throughout the entire period of study, the patients continued their usual dialysis regimen of three to five exchanges per day, At each clinic visit, a fasting blood sample (15 mL) was obtained and the serum was immediately frozen at - 4°C, The patients also brought in the outflow dialysate collected over the previous 24-hour period, Dialysate volume was calculated from the weight of the bags with and without dialysate, Aliquots, consisting of one percent of the volume of each bag, were mixed and a 7 mL sample of the pooled dialysate was stored at - 4°c' All samples were analyzed at the Wadsworth

Table 1. Composition of Peritoneal Dialysis Solutions Component (mEq/L)

Sodium Potassium Chloride Lactate Calcium Magnesium pH (unit)

Dialysate Standard

Magnesium-Free

132 0,0

132 0,0 102 35 4.0 0,0 5,3

96 40 3,5 0.5 5,2

270

SHAH ET AL

VA Medical Center. Total magnesium and calcium were measured by flameless photometry (normal range 1.6 to 2.7 mg/dL and 8.5 to lO.5 mg/dL, 0.66 to 1.1 rnrnollL and 2.1 to 2.6 rnrnoIlL); ultrafilterable magnesium and calcium by equilibrium dialysis; and ionized calcium by a specific calcium electrode. Blood and dialysate samples were collected at the end of phase I and every 2 weeks during phases II and III. Aside from the modification in the phosphate binders during phase III, other medications were continued as usual. Fifteen patients were taking supplements of vitamin D (three dihydrotachysterol and 12 calcitriol) and nine of these were receiving calcium carbonate (average 1.0 g/d). The doses of these medications were maintained constant except for one patient who required more vitamin 0 because of a preexisting tendency to develop hypocalcemia. The peritoneal fluxes of magnesium and calcium were calculated using the following equation : Flux (mg/d)

=

(Di x Vi) - (Do x Vo)

where Di and Do are the dialysate cation concentration in the inflow and outflow (mg/mL) and Vi and Vo are the dialysate inflow and outflow volumes (rnL). Negative and positive values reflect net peritoneal removal from and addition to the patient of the cations. All data are expressed as mean ± standard error. Results were compared between phases I and II and phases II and III using analysis of variance (ANOVA) for repetitive measurements. The last measurement of phase II served as the baseline for phase III. The relationship between two variables was examined by correlation analysis for linear functions. A P value < .05 was considered statistically significant.

RESULTS

Twenty-five patients entered phase II, and 22 completed the study. Three were withdrawn; two had developed mild hypomagnesemia and one inadvertently was given magnesium-containing drugs. Nine of the 22 patients who completed phase II entered phase III, the period with magnesium supplements. Table 2.

Serum (mg/dl)

Table 2 shows the results of serum magnesium (total and ultrafilterable) and calcium (total, ultrafilterable, and ionized) and net peritoneal fluxes of magnesium and calcium during each period. Serum magnesium at the end of the phase I control period was 2.2 ± 0.1 mg/dL (0.9 ± 0.04 mmoll L) . Eighteen of the patients had normal levels (1.6 to 2.7 mg/dL, 0.66 to 1.1 mmoIlL), six had mild hypermagnesemia (2.8 to 3.0 mg/dL, 1.2 mmoll L) and one had mild hypomagnesemia (1.3 mg/ dL, 0.5 mmol/L). After use of the magnesium-free dialysate was initiated, there was a significant decline in the serum magnesium level by week 2 to 1.9 ± 0.1 mg/dL (0.8 ± 0.04 mmollL) (P < .02), with subsequent stabilization at 1.8 ± 0.1 mg/dL (0.7 ± 0.04 mmol/L) throughout the 12week period (Fig 1). None of the patients, including the two whose serum magnesium level was less than 1.3 mg/dL (0.5 mmol/L), exhibited clinical signs of hypomagnesemia. During both phase I and II, serum ultrafilterable magnesium was 77 % of the total magnesium concentration, with the total and ultrafilterable values strongly correlated (r = 0.85, P < .05). There was a negative correlation (r = - 0.62, P < .05) between the peritoneal magnesium removal rate and the serum magnesium level. Peritoneal magnesium removal averaged 66 mg/d (2.8 mmol/d) at the end of phase I, and increased by 26% to 83 mg/d (3.5 mmol/d) after 2 weeks on magnesium-free dialysate (P < .05). During the remainder of phase II, magnesium removal stabilized at 73 mg/d (3.0 mmolld), with an individual range of21 to 186 mg/d (0.9 to 7.8 mmol/d). At the end of phase I, serum calcium was nor-

Serum and Dialysate Values During Control (Standard dialysate; phase I) and Magnsium-Free Dialysis (phase II) Standard Baseline

Magnesium Total 2.2 ± Ultrafilterable 1.7 ± Calcium Total 9.5 ± Ultrafilterable 5.9 ± Ionized 4.8 ± Peritoneal flux (mg/d) -66 ± Magnesium Calcium 79 ±

Magnesium-Free Dialysate Week 2

Week 4

Week 6

Week 8

Week 10

Week 12

0.1 0.1

1.9 ± 0.1 1.4 ± 0.1

1.8 ± 0.1 1.3± 0.1

1.8 ± 0.1 1.4 ± 0.1

1.8 ± 0.1 1.4 ± 0.1

1.7 ± 0.1 1.4 ± 0 .1

1.8 ± 0.1 1.4 ± 0.1

0.3 0 .2 0 .2

9.6 ± 0.3 5.8 ± 0.2 4.6 ± 0.2

9.1 ± 0.4 5.6 ± 0.2 4.4 ± 0 .2

9.6 ± 0.3 5.9 ± 0.2 4.8 ± 0.2

9.2 ± 0.2 5.9 ± 0.2 4.8 ± 0.2

8.9 ± 0.3 6.0 ± 0.2 5.0 ± 0 .2

8.9 ± 0.3 5.8 ± 0.2 5.0 ± 0.2

8 31

-83 ± 7 92 ± 23

-75 ± 5 90 ± 22

-75 ± 5 104 ± 22

-75 ± 5 102 ± 31

-67 ± 5 117 ± 25

-73 ± 6 95 ± 36

NOTE: Negative and positive flux values represent total peritoneal removal and addition, respectively. SI conversion factors: magnesium = 0.4114 mmol/L or 0.04114 mmol/d, calcium = 0.2495 mmol/L or 0.02495 mmol/d .

271

MAGNESIUM-FREE DIALYSIS IN CAPO PATIENTS

DIAL YSA TE MAGNESIUM

I~g;5

I

I

Zero

~,,"~/._lr...' - - - - - - - - - - - - - - - l .

4.0

:i '0

0, 3.0

.5 ;,:

::J

Cii ~

oct

2.0

;,:

~W

1.0

(J)

Fig 1. Serum magnesium level with dialysate magnesium concentration of 0.5 (phase I) or zero (phase II) mEq/L. Dashed lines encompass the normal range.

O~r7ll~'-r--r---'--~--'I---r---'

-4

2

4

6

10

8

phase I. Both the mean total protein and albumin levels remained at 6.3 and 3.6 g/dL (63 and 56 gl L), respectively, throughout the study period. The ultrafilterable and ionized fractions of the total serum calcium also remained constant at 60% to 68 % and 48 % to 56 %, respectively. The peritoneal calcium flux was positive and increased slightly, but not significantly, from 79 to 100 mg/d (2.0 to 2.5 mmol/d) (P > .05) . During phase III (Table 3, Fig 2), nine patients received magnesium-aluminum phosphate binders, which provided an additional magnesium load of 0.28 to 3.0 g/d. The average serum magnesium level increased from 1.8 to 2.5 mg/dL (0.7 to 1.0 mmollL) (P < .05); two patients had serum magnesium levels greater than 2.7 mgl dL (1. 1 mmol/L) (3.1 and 3.4 mg/dL, 1.3 and 1.4 mmoll

Serum and Dialysate Values During the Phase With Oral Magnesium Supplements (III) Magnesium-Free Dialysate Oral Magnesium Serum (mg/dl)

Magnesium Total Ultrafilterable Calcium Total Ultrafilterable Ionized Peritoneal flux (mg/d) Magnesium

12

WEEKS

mal in 17 (8.5 to 10.5 mg/dL, 2.1 to 2.6 mmollL), elevated in four (10.7 to 12.5 mg/dL, 2.7 to 3.1 mmollL), and reduced in four patients (6.8 to 8.1 mg/dL, 1.7 to 2.0 mmollL) . During phase II, the four hypercalcemic patients became normocalcemic, while hypocalcemia persisted in the four patients with low calcium levels. Of the 17 normocalcemic patients, four became transiently hypercalcemic (10.6 to 12.3 mg/dL, 2.6 to 3.1 mmol/L) and four developed asymptomatic hypocalcemia (5.1 to 8.4 mg/dL, 1.3 to 2.1 mmollL) during phase II. As shown in Table 2, mean total, ultrafilterable and ionized calcium levels in the serum did not show significant change (P > .05) during phase II, even though the dialysate calcium concentration was 4.0 mEq/L (2.0 mmollL) in phase II compared with 3.5 mEq/L (1.75 mmollL) in Table 3.

o

Baseline

Week 2

Week 4

1.8 1.4

± 0.1 ± 0.1

2.5 2.0

± 0.2 ± 0.2

2.5 2.1

± 0.2 ± 0.2

9.3 5.7 5.0

± 0.5 ± 0.3 ± 0 .4

9.8 ± 0.5 6.5 ± 0.4 5.4 ± 0.4

10.4 6 .3 5.0

± 0.4 ± 0.2 ± 0.4

-91

± 12

-128

±

17

-119

± 13

272

SHAH ET AL

e

180 ~

>-

01 "C

:i "C

""-

""-

co

co 150 .§

5

E

W

~

I-

(/)

:::>

II:

0

:z:

0..

(/)

0

:z:

0..

I

0

z <

< II: ..J 120 <

4

>

o

::,;

w 90

3

;2 ..!J (/)~ w

::,;

z

:::>

iii e w z I Cl 9 < ::,;

.

::,; ~

II:

2

80

~

::,; ..J

< w

::,;

30

:::>

II:

~

l-

ii:

w

w

(/)

0

n

m

L), but without clinical signs of magnesium toxicity. The failure of serum magnesium levels to rise to higher values may be explained, in part, by the concomitant increase in peritoneal magnesium removal from 91 to 128 mg/d (3 .8 to 5.3 mmol/d) (P < .02). The efficacy of the magnesium-aluminum binders in controlling serum phosphorus is shown in Fig 2 for the seven patients in whom data were available. Even though the dose of aluminum containing phosphate binders was reduced, the serum phosphorus level did not rise . In fact, it fell slightly but not significantly from 5.4 ± 0.6 mgl dL (1.7 ± 0.2 mmol/L) at the end of phase II to 4.8 ± 0.6 mg/dL (1 .5 ± 0.2 mmollL) at 2 and 4 weeks in phase III (P > .05). The altered dialysate composition (Table I) had an effect on the patients' acid-base status. Serum total carbon dioxide content decreased from the phase I control value of 24 ± 0.8 mEq/L to 20.1 ± 1.0 mEq/L (mmollL) by week 2 of the magnesium-free dialysate period (P < .05), and remained at this level for the remainder of the study. Conversely, there was a rise in the serum chloride concentration from 97.6 ± 0.9 mEq/L (mmollL) to 102.0 ± 0.8 mEq/L (mmoI/L) during this same period (P < .05) . The mild metabolic acidosis observed during phase II did not have a significant effect on the serum level of ionized calcium (P > .05) . There were no changes in serum sodium, potassium, urea nitrogen , or creatinine.

o

0..

Fig 2. Serum magnesium and phosphorus levels and peritoneal magnesium removal rate during the periods with standard (I) and magnesium-free dialysate, the latter with (III) or without (II) oral magnesium supplementation.

DISCUSSION

The objective of this study was to assess the effects of a magnesium-free dialysate on certain aspects of magnesium metabolism in patients on CAPD. Specifically, we sought to determine net magnesium removal and the resultant effects on serum magnesium levels . In addition, we evaluated the short-term effects of administration of oral magnesium-containing phosphate binders on the serum magnesium concentration and peritoneal flux of magnesium. Clinical consequences of these manipulations were also monitored. At the end of the control period, which represented at least 3 months of treatment using a standard dialysate concentration of magnesium of 0.5 mEq/L (0.25 mmol/L), the mean serum magnesium concentration remained within normal limits (2.2 mg/dL, 0 .9 mmollL); these data are similar to the observations of Nolph et al. I9 During the early part of the magnesium-free dialysate phase, serum magnesium fell by 14% and remained stable thereafter. This decrement is partly due to the 26 % increase in peritoneal magnesium efflux observed during the same time period. The subsequent return of peritoneal efflux toward the baseline value could be attributed to the new, but lower, steadystate serum magnesium level. Panarello et al 20 also studied a magnesium-free dialysate over a 4 week period. In their preliminary report, the serum magnesium level fell to 1.8 mg/dL (0.7 mmol/L)

273

MAGNESIUM-FREE DIALYSIS IN CAPO PATIENTS 3.5

80

70

:::; ~ co

oS z

80

..

zw

60

~

40

o

~

~ co

..oS

w

c

a:

-'

c a:

c

~O o

>

0

::E w

~O

::>

iii

:2

rn w

z c

w

Cl

CI

~

z

c

Fig 3. Summary data from several studlesI8.21 .pr •••nt study showln'g average serum magnesium level and peritoneal removal of magnesium as a function of dialysate magnesium concentration.

.

~

30

~

::E ::>

-'

cw

.. Z

0

a:

w

rn

20

it

...w

0 0 .9

1.5

0.5

0

DIALYSATE MAGNESIUM CONCENTRATION (mEq/L)

and peritoneal removal rate increased to 79 mg/d, values that are similar to those in the present study. That the magnesium concentrations in the serum and dialysate and the peritoneal removal rate are interrelated is supported by other studies in addition to our own observations. With a dialysate magnesium concentration of 1.5 mEq/L (0.75 mmollL), the peritoneal removal rate is 46 mg/d (1.9 mmol/d)y·ls As the dialysate concentration is progressively reduced to 0.9,0.5, and 0 .0 mEq/ L (0.45, 0.25, and 0.0 mmol/L), the peritoneal removal increases to 57 (2.4),21 66 (present study) and 75 mg/d (3.1 mmolld) (present study, 20, 2.8). As shown in Fig 3, the serum magnesium levels corresponding to these dialysate concentrations and peritoneal removal rates are 3.0, 2.4, 2.2, and 1.8 mg/dL (1.2, 1.0,0.9, and 0.7 mmoll L). Since these data are drawn from different studies, it is not possible to do a correlation analysis. However, it appears that decreasing dialysate concentration of magnesium is associated with enhanced peritoneal removal rate, and this probably accounts for the concomitant fall in serum magnesium concentration. This cross-sectional evaluation of data from the literature does not allow us to speculate about a direct effect of serum magnesium concentration on peritoneal removal rate, since different dialysate concentrations were used. However, our longitudinal study using a magnesium-free dialysate suggests that the serum level may indeed affect peritoneal flux . Obviously,

careful balance studies are needed to define the interrelationship of these variables. When patients were administered the magnesium-aluminum phosphate binders in phase III, the serum magnesium level increased by 39 %, and this was associated with a 41 % increase in peritoneal magnesium removal (Fig 2) . The percent change in serum magnesium level was related to the dose of oral magnesium supplement (r = 0.66, P < .05). In addition, there was a positive correlation between the percent increase in peritoneal magnesium removal rate and the percent increase in serum magnesium level (r = 0.77, P < .02). These observations suggest that the rise in serum magnesium level resulting from the oral supplement was minimized by the enhanced peritoneal removal rate. In fact, only two of these nine patients developed mild and asymptomatic hypermagnesemia , as the magnesium-free dialysate was used. This compares with elevated serum magnesium levels (3.2 and 4.3 mg/dL, 1.3 and 1.8 mmol/L) in two of four patients who had received similar magnesium supplements during treatment with dialysate containing 0.5 mEq/L (0.25 mmoll L) of magnesium (unpublished observations, April, 1984). However, since this was not a balance study, other factors such as dietary intake, tissue distribution, and fecal losses could have contributed to these changes. It should be noted that the phosphate binder therapy combining magnesium and aluminum in

274

SHAH ET AL

phase III, reduced the oral aluminum burden by 22 % to 75 %, but still effectively controlled the serum phosphorus level (Fig 2). Our results are similar to the observations of O'Donovan et al and Guillot et al in hemodialysis patients. 15.16 During the entire 16 week period of magnesiumfree dialysate (phase II and III), neither serum calcium level nor peritoneal calcium flux changed, even though the dialysate calcium concentration was 0 .5 mEq/L (0.25 mmollL) greater than that in the control period. Calcium metabolism has multiple determinants and our data do not permit an analysis of the effects of other factors, such as baseline serum calcium levels, calcium and vitamin D ingestion and absorption, and the degree of uremic bone disease. Also, it has been shown that in experimental magnesium depletion, the equilibrium of calcium between the extracellular fluid and bone is altered. 22 The mild metabolic acidosis that developed in our patients is similar to that reported for other CAPD patients when the dialysate lactate concentration is 35 mEq/L (mmoIlL).17 It has been shown

that this metabolic abnormality can be corrected by increasing the lactate concentration to 40 mEq/ L (mmoIlL).19 Thus, we have shown that, in general, the serum magnesium level remains within the normal range during peritoneal dialysis using a magnesium-free dialysate if the baseline value is not subnormal. With this regimen , oral magnesium-aluminum phosphate binders can safely and effectively be administered. Such a treatment regimen can reduce the oral load of aluminum, which might have a beneficial effect on the development of serious complications attributed to excess tissue aluminum, such as dialysis encephalopathy and osteomalacia. Further studies are needed to determine the long-term effects of this regimen of dialysate and phosphate binder on aluminum, phosphorus, and magnesium metabolism. ACKNOWLEDGMENT We express our gratitude to the dialysis staff at the participating institutions for their help and to Dr Daniel Schwartz for statistical assistance.

REFERENCES I. Massry SG, Sellig MS : Hypomagnesemia and hypermagnesemia. Clin Nephrol 7: 147-153 , 1977 2. Contiguglia SR, Alfrey AC, Miller NL, et a\: Total body magnesium excess in chronic renal failure . Lancet I : \300\302, 1972 3. Contiguglia SR, Alfrey AC, Miller NL, et al: Nature of soft tissue calcification in uremia. Kidney In! 4:229-235, 1973 4. Randall RE, Cohen MD, Spray CC, et al: Hypermagnesemia in renal failure, etiology and toxic manifestations. Ann Intern Med 61:73-88, 1964 5 . Catto GRD, Reid IW, MacLeod M: The effect of low magnesium dialysate on plasma, ultrafilterable, erythrocyte and bone magnesium concentrations from patients on maintenance hemodialysis. Nephron 13:372-381, 1974 6. Gonella M: Plasma and tissue levels of magnesium in chronically hemodialyzed patients: Effects of dialysate magnesium levels. Nephron 34:141-145 , 1983 7. Alfrey AC, Miller NL: Bone magnesium pools in uremia. J Clin Invest 52:3019-3027, 1973 8. Gonella M, Moriconi L, Betti G, et al: Serum levels of PTH, Mg, Ca, inorganic phosphorous and alkaline phosphatase in uremic patients on differentiated Mg dialysis. Proc Eur Dial Transplant Assoc 17:362-366, 1980 9. Massry SG, Coburn JW, Hartenbower DL, et al: Mineral content of human skin in uremia. Effect of secondary hyperparathyroidism and hemodialysis. Proc Eur Dial Transplant Assoc 7: 146-153 , 1970 10. Fleming LW, Lenman JAR , Stewart WK: Effect of magnesium on nerve conduction velocity during regular dialysis treatment. J Neurol Neurosurg Psychiatr 35 :342-355, 1972

II. Graf H, Kovarik J, Stummvoll H, et al: Disappearance of uremic pruritus after lowering dialysate magnesium concentration. Br Med J 2: 1478-1479, 1979 12. Llach F: Effects of renal failure and dialysis on divalent ion metabolism, in Brenner BM, Stein J (eds): Contemporary Issues in Nephrology: Divalent Ion Homeostasis. New York, Churchill Livingstone, 1983, pp 291-317 \3. Alfrey AC: Aluminum metabolism. Kidney Int 29:S8Sl1, 1986 (suppl) 14. Ott SM, Maloney NA, Coburn JW, et al: The prevalence of bone aluminum deposition in renal osteodystrophy. N Engl J Med 307:709-713, 1982 15. O'Donovan R, Hammer M, Baldwin D, et al: Substitution of aluminum salts by magnesium salts in control of dialysis hyperphosphatemia. Lancet 1:880-883 , 1986 16. Guillot Ap, Hood VL, Runge CF, et al: The use of magnesium-containing phosphate binders in patients with end stage renal disease on maintenance hemodialysis. Nephron 30: 114-117, 1982 17. Nolph KD, Sorkin M, Rubin J, et al: Continuous ambulatory peritoneal dialysis: Three-year experience at one center. Ann Intern Med 92:609-6\3, 1980 18. Blumenkrantz MJ, Kopple JD , Moran JK, et al: Metabolic balance studies and dietary protein requirements in patients undergoing continuous ambulatory peritoneal dialysis. Kidney Int 21:849-861 , 1982 19. Nolph KD, Prowant B, Serkes KD, et al: Multicenter evaluation of a new peritoneal dialysis solution with a high lactate and a low magnesium concentration . Perit Dial Bull 3:6365, 1983

MAGNESIUM-FREE DIALYSIS IN CAPO PATIENTS

20. Panarello G, Schinella D, Raimonds A, et al : Evaluation of the effect of different concentrations of calcium and magnesium in solutions for CAPD. Kidney Int 24:S334, 1983 (abstr) (suppl 16)

275 21. Unpublished data: Abbott Laboratories, North Chicago, IL 22. Reddy CR, Coburn JW, Hartenbower DL, et al: Studies on mechanisms of hypocalcemia of magnesium depletion. J Clin Invest 52:3000-3010, 1972