Kidney International, Vol. 44 (1993), pp. 1139—1144

Does prolonged protein restriction preceding dialysis lead toprotein malnutrition at the onset of dialysis?

MACKENZIE WALSER

Johns Hopkins University School of Medicine, Department of Pharmacology and Molecular Sciences, and Department of Medicine,Baltimore, Maryland, USA

Does prolonged protein restriction preceding dialysis lead to proteinmalnutrition at the onset of dialysis? It has recently been suggested thatprolonged protein restriction preceding dialysis may induce proteinmalnutrition and thus confer a poor prognosis during dialysis. Weexamined the records of all patients who were prescribed a very lowprotein diet (0.3 g/kg ideal body weight) plus supplemental essentialamino acids and/or ketoacids for 6 to 72 months (median 26 months)preceding renal replacement, numbering 43. Hypoalbuminemia imme-diately preceding dialysis was present in only two patients. Final serumalbumin averaged 4.1 0.4 (SD) g/dl. Final transferrin was subnormal ineight patients, but had been subnormal in six of these, without attendanthypoalbuminemia, for one to four years. Mean final transferrin was 241

56 mg/dl. Final serum cholesterol was below 150 mg/dl in sixsubjects, all of whom had normal levels of albumin (mean 4.0 0.2 g/dl)and a normal mean value for transferrin (211 22 mg/dl). In fivepatients who exhibited subnormal albumin and transferrin and highcholesterol concentrations at the beginning of dietary therapy, albuminand transferrin levels rose to normal or nearly normal, and hypercho-lesterolemia receded during the ensuing four months. Thus this pre-dialysis dietary regimen, rather than causing protein malnutrition,prevents it; when protein malnutrition is present, this regimen corrects it.

Even though protein-restricted diets have been known fordecades to reduce symptoms in chronic renal failure, as re-viewed by Bergstrom [1], their utilization in such patients is farfrom universal. One reason is concern that prolonged consump-tion of a low protein diet may cause protein malnutrition, asrecently pointed out by Hakim and Levin [2]. 1-lypoalbumine-mia is now known to be the single greatest risk factor for deathduring dialysis [3—7].

Having followed patients on very low protein diets supple-mented by essential amino acids or ketoacids for some years,we now have records on 43 who have gone on to renalreplacement therapy. We therefore undertook to documenttheir protein nutrition at the termination of nutritional therapyand beginning of dialysis or transplantation. We also examinedthe records of five patients who exhibited protein malnutritionbefore the start of nutritional therapy. The results show thatrather than causing protein malnutrition, this regimen preventsit and can also correct it.

Received for publication April 6, 1993and in revised form June 24, 1993Accepted for publication June 24, 1993

© 1993 by the International Society of Nephrology

Methods

Records of all 43 patients who were prescribed a very lowprotein (0.3 glkg ideal body weight) diet plus supplementalamino acids or ketoacids for six months or longer, who thenwent on to dialysis (41 patients) or transplantation withoutpreceding dialysis (2 patients), were examined. The decisionwhen to initiate dialysis was made by the director of the dialysisunit, using the same clinical criteria as were applied to patientsnot on dietary treatment.

Initial serum creatinine concentration was 4.8 1.6 mgldl.Results of sequential determinations of glomerular filtration rateand other measurements in most of these patients duringnutritional therapy appear in previous reports from this labora-tory [8—12]. Mean values throughout dietary therapy werecalculated for each patient and averaged across patients for17-hydroxycorticoid excretion, free cortisol excretion, urea Nexcretion, serum albumin, serum transferrin, and serum CO2.Final values immediately preceding renal replacement weresummarized for serum albumin, transferrin, cholesterol, urea Nand creatinine. Since most patients were given amino acid orketoacid supplements alternately, no attempt was made todifferentiate between the number of months during which eachof these two supplements was consumed.

The diet has been described in detail elsewhere, as has thecomposition of the supplements and the use of multivitaminsand CaCO3 [8—12]. Patients were seen 5 to 12 times a year bythe author, and by the dietician several times at the outset andthereafter on request of patient or physician. Caloric prescrip-tion was generally 35 kcallkg ideal body weight, but was lowerin subjects who desired to lose weight. Calorie counts duringtreatment were performed only when undesirable weight lossoccurred. Only one patient developed hypoalbuminemia duringtreatment (other than the two patients who developed hy-poalbuminemia just before dialysis, as detailed in Results).Supplement dosage was doubled and serum albumin returned tonormal.

Five patients who were hypoalbuminemic when first seenwere reviewed in greater detail. Before treatment, proteinuriain these patients averaged 7.1 4.2 glday. Four of these areamong the 43 included in the preceding analysis; the fifth wasexcluded because he completed fewer than six months ofnutritional therapy before going on dialysis. These five patientswere treated with the same regimen, except that three of themreceived 20 g per day of essential amino acids instead of 10 g per

1139

1140 Walser: Does protein restriction lead to malnutrition7

day. Values for serum albumin, transferrin, cholesterol and24-hour protein excretion were measured monthly and averagedacross patients.

Prealbumin was not measured. Creatinine-height index wasnot used because, even when corrected for age, it is a highlyunreliable measure of muscle mass in the presence of chronicrenal failure [131. Low values are commonly seen at GFR'sbelow 25 mllmin, but the roles of decreased protein intake,increased creatjnine metabolism and reduced muscle mass incausing these low values are uncertain [14].

All measurements were performed by the Johns HopkinsHospital Chemistry Laboratory, using automated methods.

Means and standard deviations are reported. Changes innutritional parameters were evaluated by Student's t-test forpaired samples.

Results

Follow-up periodsThe distribution of follow-up times is shown in Figure 1. The

median was 26 months.

Final albumin concentrationsThe distribution of final serum albumin levels is depicted in

Figure 2. Only two patients had values below 3.4 gldl. Both ofthem became hypoalbuminemic in the final month precedingdialysis, during which they were anorexic and vomiting. Inretrospect, they should probably have been started on dialysisearlier.

Final serum albumin was not correlated with the duration ofprotein restriction (Table 1).

Final transferrin concentrationsFinal serum transferrin values are depicted in Figure 3. Eight

patients had levels below 200 mg/dl, including the two hypoal-buminernic patients. In the other six patients, examination ofearlier values for transferrin and albumin reveals that four ofthem had exhibited similar values for transferrin for one to fouryears, without hypoalbuminemia. The remaining two patientsdeveloped subnormal levels of transferrin only in the finalmonths. Even though albumin levels remained normal, it isconceivable that the decrease in transferrin in these two pa-tients was a sign of impending protein deficiency.

Final cholesterol concentrations

Final values for serum cholesterol are depicted in Figure 4.Six patients had levels below 150 mg/dl, all of whom had normallevels of albumin (mean 4.0 0.2g/dl) and a normal mean valuefor transferrin (211 22 mg/dl). Cholesterol concentration wassignificantly (P < 0.05) correlated with transferrin concentra-tions (Table 1).

Other measurements

Final serum urea N was 82 24 mg/dl. Final serum creatininewas 8.8 2.3 rng/dl. These values were significantly correlated,as expected, but were not correlated with the nutritionalparameters except that final serum urea N was correlated withurinary urea N (Table 1).

Body mass index (weightlheight2) averaged 24 4 kg/rn2.Body weight decreased by 0.14 0.26 kg/month (P < 0.001),

but a number of patients (for example, Fig. 5 in [10]) voluntarilyrestricted their caloric intake in order to lose weight. Weightloss was not correlated with nutritional parameters. Averageserum albumin during treatment (4.2 0.3 g/dl) was slightlyhigher (by 0.12 0.32 g/dl, P <0.02) than final serum albumin.The average value during treatment for serum transferrin wasnot significantly different from the final value, 241 56 mg/dl.Average serum CO2 during treatment was 24 4 mri. Averageurine urea N was 5.0 1.2 g/day (range 2.9 to 8.3 g/day),indicating an average estimated protein intake of 34 g (thesupplements contained 1.4 to 1.8 g N). Protein intake estimatedin this way was significantly correlated with final serum urea Nand with 17-hydroxycorticoid excretion (as we have reported

18

17

16

15

14

13

12

11

8

7

6

5

4

3

>'C)

a)

a)U-

12 24 36 48 60 72

Months midpointFig. 1. Distribution of follow-up times on diet (months). "Midpoint"means midpoint of range for each vertical bar. Hence the first barsignifies 6 to 18 months on diet (patients followed less than 6 monthswere excluded).

2

1

0

SUN Cr AIb Trans Chol 170H Coil UNA CO2

Mos NS NS NS NS NS NS NS NS NSSUN 0.38a NS NS NS NS NS O.30c NSCr NS NS NS NS NS NS NSAib Ø•47b NS NS NS NS NSTrans 0.32c NS NS NS NSChol NS NS NS NS170H 041b 041b NSCoil NS NSUNA NS

Abbreviations are: Mos, months of protein restriction; SUN, finalserum urea N; Cr, final serum creatinine; Aib, final serum albumin;Trans, final serum transferrin; Chol, final serum cholesterol; 170H,average 17-hydroxycorticoid excretion during diet; Coil, average freecortisol excretion during diet; UNA, average urinary urea N excretionduring diet; CO>, average serum CO2 during diet; NS, not significant.

ap <0.02b P < 0.01c P < 0.05

11

>.C.)

a,

a>U-

10

9.

8-

7.

6.

5

4

3.

2

1

0

Walser: Does protein restriction lead to malnutrition? 1141

30

20

a)

a>U-

10

0

3.15 3.45 3.75 4.05 4.35 4.65 4.95Albumin midpoint

Fig. 2. Distribution offinal values for serum albumin, gidi. Midpoint asdefined in Figure 1 legend.

Table 1. Correlation coefficients between measured parameters

150 210 270 330 390 450

Transferrin midpointFig. 3. Distribution of final values for serum transferrin,mg/dl. Mid-point as defined in Figure 1 legend.

previously in [91) but not with other parameters (Table 1).17-Hydroxycorticoid excretion averaged 3.3 1.1 mg/day(normal 3 to 9 mg/day). Free cortisol excretion averaged 24 4pg/day (normal 25 to 125 gIday).

Sequential data in five hypoalbuminemic patientsAt the start of protein restriction, mean values for serum

albumin, transferrin and cholesterol in these five patients were3.1 0.3 g/dl, 208 36 mg/dl, and 282 80 mg/dl, respectively.During four months of nutritional therapy albumin levels rose tonormal in four of the five patients. The final mean was 3.5 0,4g/d1. Transferrin also rose to normal in four of the five patients;the final mean was 242 28 mg/dl. Serum cholesterol fell from282 80 to 205 43 mg/dl. The significance of paired changesin all three parameters is shown in Figure 5. Urinary protein

1142 Walser: Does protein restriction lead to malnutrition?

Cholesterol midpointFig. 4. Distribution of final values for serum cholesterol, mg/dl. Mid-point as defined in Figure 1 legend.

excretion decreased 32%, but the change was not statisticallysignificant.

Discussion

In these 43 patients followed 6 to 72 months on severe proteinrestriction plus supplemental amino acids or ketoacids, signs ofprotein malnutrition at the onset of renal replacement are rare.At the most, four patients may have been (or were about tobecome) protein-deficient. In two of these, anorexia and vom-iting during the last month of follow-up was the probable cause.Thus, protein malnutrition was clearly not induced by thisregimen.

In contrast, hypoalbuminemia is seen in 25 to 50% of patientsentering dialysis nationwide [3—7]. How many of these wererestricting protein intake prior to dialysis is conjectural, but it

Time, months

Fig. 5. Sequential observations during nutritional therapy in five pa-tients who were hypoalbuminenic at the start. Shaded areas representpart of the normal range for each parameter. Significance of changesfrom time zero indicated as follows: * < 0.01; < 0.05.

may be safely assumed that very few were prescribed thesupplemented low protein diet utilized in the present study.

Long-term observation of pre-dialysis patients following reg-imens identical or similar to the present regimen has generallyfound no signs of protein malnutrition [8—22], with one excep-tion [23] in which protein intake was more severely restricted(to 0.2 glkg). The results of different measures of proteinnutrition in this latter study [23] were inconsistent. None ofthese studies reports final values for nutritional parameters atthe onset of dialysis.

If replacement of a substantial portion of dietary protein byessential amino acids or ketoacids improves protein nutrition,as these results indicate, the question arises as to whether someconstituent of proteinaceous foods contributes to protein mal-nutrition. Accumulation of uremic toxins derived from foodprotein may contribute to anorexia, thereby leading to inade-quate energy intake. Speculatively, glucocorticoids may play a

E

18

17

16

15

14

13

12

11

10

98

76

5.

4.

a)

a)U-

300

200

100

300

200

100

4

3

2

Transferrin

E

I I I I I

0

Albumin

I I

120 160 200 240 280 320

1 2 3 4

Walser: Does protein restriction lead to malnutrition? 1143

role. It is clear that intake of whole protein stimulates glucocor-ticoid production [24—26, Table 1]. High physiological levels ofglucocorticoids can increase protein breakdown [27, 28] andcause negative N balance [29, 30]. No comparison of pureessential amino acids with whole protein in their propensity tostimulate glucocorticoid production has apparently been re-ported, but it is possible that less stimulation occurs with aminoacids. Steroid excretion rates in the present subjects weregenerally low, as has been reported previously in pre-dialysispatients [9, 31].

The five hypoalbuminemic patients studied were not severelynephrotic (mean proteinuria 7.1 g/day), but it is likely thatproteinuria was the cause of their protein deficiency. Theoptimal diet to improve hypoalbuminemia in nephrotic patientsremains uncertain. Protein restriction is known to reduce pro-teinuria [32], perhaps by reducing glucocorticoid production[24—26], which increases proteinuria [33—35]. Nephrotic patientsmay benefit from protein restriction plus supplementation withessential amino acids and/or ketoacids in terms of proteinuriaand hyperlipidemia, but no improvement in hypoalbuminemiahas been found [36, 37]. Whether double doses of essentialamino acids plus protein restriction can improve hypoalbumin-emia in frankly nephrotic patients remains to be determined.

Acknowledgments

This work was supported by a Research Grant (DK-32008) and aClinical Center Grant (RR-00722) from the National Institutes ofHealth.

Reprint requests to Mackenzie Walser, M.D., Johns Hopkins Schoolof Medicine, 725 N. Wolfe Street, Baltimore, Maryland 21205, USA.

References

1. BERGSTROM, J: Discovery and rediscovery of low protein diet. ClinNephrol 21:29—34, 1984

2. HAKIM RM, LEvIN N: Malnutrition in hemodialysis patients. Am JKidney Dis 21:125—137, 1993

3. GOLDWASSER P, MITTMAN N, ANTIGNANI A, BURRELL D, MICHELM-A, COLLIER J, AVRAM MM: Predictors of mortality in hemodi-alysis patients. JAm Soc Nephrol 3:1613—1622, 1993

4. LOWRIE EG, LEw NL: Death risk in hemodialysis patients: Thepredictive value of commonly measured variables and an evalua-tion of the death rate differences between facilities. Am J KidneyDis 15:458—482, 1990

5. ANONYMOUS: Comorbid conditions and correlations with mortalityrisk among 3,399 incident hemodialysis patients. Am J Kidney Dis20 (Suppl 2):32—38, 1992

6. KAMINSKI MV JR, L0wRIE EG, HAA5E T: Malnutrition is lethal,diagnosable, and treatable in ESRD patients. Transplant Proc23:1810—1815, 1991

7. CHURCHILL CN, TAYLOR DW, COOK RJ, LAPLANTE P, BARRE P,CARTIER P, FAY WP, GOLDSTEIN MB, JINDAL K, MANDIN H,MCKENZIE JK, MUIRHEAD N, PARFREY PS, P05EN GA, SLAUGH-TER D, ULAN RA, WERB R: Canadian hemodialysis morbiditystudy. Am J Kidney Dis 19:214—234, 1992

8. WALSER M, LAFRANCE ND, WARD L, VANDUYN MA: Progressionof chronic renal failure in patients given ketoacids following aminoacids. Kidney mt 32:123—128, 1987

9. WALsER M, WARD L: Progression of chronic renal failure is relatedto glucocorticoid production. Kidney mt 34:859—866, 1988

10. WAL5ER M: Progression of chronic renal failure in man. Kidney mt37:1195—1210, 1990

11. WALsER M, HILL S, WARD L: Progression of chronic renal failureon substituting a ketoacid supplement for an amino acid supple-ment. JAm Soc Nephrol 2:1178-1185, 1992

12. WAL5ER M, HILL S, WARD L, MAGDER L: A crossover comparisonof progression of chronic renal failure: Ketoacids vs. amino acids.Kidney mt 43:933—939, 1993

13. WALSER M: Creatinine excretion as a measure of protein nutritionin adults of varying age. J Par Ent Nutr 11 :73S—78S, 1987

14. Modification of Diet in Renal Disease (MDRD) Study Group.Prepared by KOPPLE JD, BERG R, HOUSER H, STEINMAN TI,TESCHAN P: Nutritional status of patients with different levels ofchronic renal insufficiency. Kidney mt 36 (Suppl 27):S184-.S194,1989

15. TEPLAN V, SCHUCK 0, NADvORNIKOVA H, RUZICKOVA J,GRAFNETTER D, KASLIK J: Stoffwechselverhalten bei Patienten mitChronischer Nierensuffizienz unter Diatëtischer Langzeittherapieund Substitution mit Ketoanalogen Essentieller Aminosäuren. ZUrol Nephrol 83:1989—1996, 1990

16. SCHMICKER R, WETTER K, LINDENAU K, FROHLING PT, KOKOTF: Conservative long-term treatment of chronic renal failure withketo acid and amino acid supplementation. mnfusionstherapie 14(Suppl 5):34—38, 1987

17. MEI5INGER E, STRAUCH M: The influence of two different essentialamino acid/keto analogue preparations on the clinical status ofpatients with chronic renal failure. Zeit Ernahrungswissenschaft24:96—104, 1985

18. ALVESTRAND A, FURST P, BERGSTROM J: Plasma and muscle freeamino acids in uremia: Influence of nutrition with amino acids. ClinNephrol 18:297—305, 1982

19. ALVESTRAND A, AHLBERG M, FURST P, BERGSTROM J: Clinicalresults of long-term treatment with a low-protein diet and a newamino acid preparation in patients with chronic uremia. ClinNephrol 19:67—73, 1983

20. ATTMAN P-O, EwALD J, I5AKSSON B: Body composition duringlong-term treatment of uremia with amino acid supplementedlow-protein diet. Am J Clin Nutr 33:801—810, 1980

21. NOREE L-O, BERGSTROM J: Treatment of chronic uremic patientswith protein-poor diet and oral supply of essential amino acids. II.Clinical results of long-term treatment. C/in Nephrol 3:195—203,1975

22. MARIANI G, BARSOTTI G, CIARDELLA F, MOLEA N, MORELLI E,NAZZUCA N, Niosi F, BONAGUIDI F, FuSANI L, PANICUCCI F,GIOVANNETTI S, BIANCHI R: Albumin metabolism and nutritionalstatus of uremic patients on a long-term very-low-protein dietsupplemented with essential amino acids and keto analogues. JNuci Med Allied Sci 28:237—244, 1984

23. LUCAS PA, MEADOWS JH, ROBERTS DE, COLES GA: The risks andbenefits of a low protein-essential amino acid-keto acid diet. Kidneymt 29:995—1003, 1986

24. FOLLENIUS M, BRANDERBERGER G, HETTER B: Diurnal cortisolpeaks and their relationship to meals. J Gun Endocrin Metab55:757—761, 1982

25. RASLIWAN SS, ISMAIL NA, HAMDY MA, HASSAN HA, ABDEL-MAKSOUD AM: Effect of dietary protein and fat on the adrenocor-tical function. Nutr Reports mt 37:289—302, 1988

26. ANDERSON KE, ROSNER W, KHAN MS, NEW MI, PANG S, WISSELPS, KAPPAS A: Diet-hormone interactions: Protein/carbohydrateratio alters reciprocally to plasma levels of testosterone and cortisoland their respective binding globulins in man. Life Sci 40:1761—1768, 1987

27. QUAN ZY, WALSER M: The effect of corticosterone administrationat varying levels on leucine oxidation and whole body proteinsynthesis and breakdown in adrenalectomized rats. Metabolism40:1263—1267, 1991

28. DARMAUN D, MATTHEWS DE, BIER DM: Physiological hypercorti-solemia increases proteolysis, glutamine, and alanine production.Am J Physiol 255:E366—E373, 1988

29. QUAN ZY, WALSER M: Effects of corticosterone administration onnitrogen excretion and nitrogen balance in adrenalectomized rats.Am J C/in Nutr 55:695—700, 1992

30. GELFAND RA, MATTHEWS DE, BIER DM, SHERWIN RS: Role ofcounterregulatory hormones in the catabolic response to stress. JC/in Invest 74:2238—2248, 1984

31. HENNING HV, LUDWIG-KOHN H: Neutral steroid metabolites inbody fluids of patients with uremia and after renal transplantation,

1144 Walser: Does protein restriction lead to malnutrition?

in Contributions to Nephrology (vol 50), edited by BERLYNE GM,GIOVANNETTI S, Basel, Karger, 1986, pp. 153—166

32. KAYSEN GA, GAMBERTOGLIO J, JIMENEZ I, JONES H, HUTCHIN-SON FN: Effect of dietary protein intake on albumin homeostasis innephrotic patients. Kidney mt 29:572—577, 1986

33. WETZELS JFM, GERLAG PGG, SLUITER HE, HOITSMA AJ, KOENERAP: Prednisone-induced fluctuations of proteinuria in patientswith a nephrotic syndrome. Nephron 44:344—350, 1986

34. PESSINA AC, PEART WS: Renin induced proteinuria and the effectsof adrenalectomy. I. Haemodynamic changes in relation to func-tion. Proc R Soc Lond B 180:43—60, 1972

35. MAEDA T, FURUKAWA I, SUZUKI T, Moai H: Reduction ofproteinuria by adrenalectomy and its restoration by glucocorticoidsin rats bearing functioning pituitary tumor. Nephron 55:181—186,1990

36. BARSOTTI G, MORELLI E, CuPIsTI A, BERTONCINI P, GI0vAN-NETTI 5: A special, supplemented "vegan" diet for nephroticpatients. Am J Nephrol 11:380—385, 1991

37. CIARDELLA F, ROMANO F, MAssrrELLI M: Hyperlipidemia innephrotic patients with chronic renal failure: Effects of a low-protein vegan diet. (abstract) J Nephrol 4:125, 1991