Treatment of Nephrotic Adults With a Supplemented,

Very Low-Protein Diet

Mackenzie Walser, MD, Sylvia Hill, BS, and Elizabeth A. Tomalis, MPH, RD

0 Optimal dietary protein intake for adults with the nephrotic syndrome has not been established; very low-
protein diets are believed to be contraindicated. Sixteen patients with the nephrotic syndrome were nevertheless
prescribed a very low protein diet (0.3 g/kg) supplemented by 10 to 20 g/d essential amino acids (or, in a few
cases, ketoacids) for an average of 10 months (range, 1 to 36 months). In 11 patients with initial glomerular
filtration rates (GFRs) 530 mUminl3 m* of height (ht)*, significant but modest improvement was seen (on the
average) in proteinuria, serum albumin, and serum cholesterol; all 11 eventually went on to dialysis. The other
five patients, with initial GFRs of 32 to 69 ml/minM m2 of ht’, had either focal segmental glomerulosclerosis,
diabetic nephropathy, or, in one patient, both. The nephrotic syndrome associated with these disorders rarely
remits spontaneously. However, during the following 3 to 15 months mean proteinuria decreased from 9.3 to 1.9
g/d, mean serum albumin increased from 2.5 g/dL to 3.6 g/dL, and mean serum cholesterol decreased from 415
mg/dL to 255 mg/dL (all P < 0.001). The GFR either remained constant or increased. Four of these five patients
have resumed normal or nearly normal diets and remain in remission or near-remission for 6 to 24 months. We
conclude that severe protein restriction plus an essential amino acid supplement may induce prolonged remission
in adults with the nephrotic syndrome provided that GFR is not severely reduced. The mechanism of this paradoxi-
cal response to protein restriction remains to be determined.
0 1996 by the National Kidney Foundation, Inc.

INDEX WORDS: Nephrosis; diet protein; serum albumin; proteinuria; serum cholesterol; glomerular filtration rate;
amino acids.

I N THE PAST, nephrotic syndrome has been

treated by a high-protein diet’ because there
was evidence that positive nitrogen balance could
generally seen.9-25 Severely protein-restricted
diets have been believed to be contraindicated
because of the danger of protein malnutrition,
be achieved in this way.2-5 However, it has long based in part on the results of studies in rats.26z27
been known that high-protein intake increases We have recently reported*’ that a supple-
proteinuria,2M5 and that proteinuria may promote mented, very low-protein diet, when adminis-
glomerulosclerosis6 and tubulointerstitial injury7 tered for at least 6 months predialysis, almost
and is correlated with the rate of progression of completely prevents hypoalbuminemia at the on-
chronic renal failure, according to many studies.* set of dialysis in patients with chronic renal fail-
Hence, the use of high-protein diets in adults ure, in contrast to a 25% to 50% incidence of
with the nephrotic syndrome has been largely hypoalbuminemia at the onset of dialysis nation-
abandoned. wide. Included in this report was a graph de-
It is still not established, however, whether a picting mean values in five patients (nos. 23, 44,
normal or moderately restricted intake of protein 46, 60, and 65 of the present report) who pre-
should be recommended. Recent studies have sented with hypoalbuminemia as well as chronic
shown that protein restriction may reduce pro- renal failure; proteinuria averaged 7 g/d. After 4
teinuria and hypercholesterolemia; however, lit- months of this diet, serum albumin had increased
tle or no improvement in hypoalbuminemia is to nearly normal and serum cholesterol had de-
creased to nearly normal; proteinuria decreased,
From the Department of Pharmacology and Molecular but not significantly. We also reportedz9 that pa-
Sciences and the Department of Medicine, Johns Hopkins tients who followed a supplemented low-protein
School of Medicine and the General Clinical Research Cen- diet for at least 6 months before starting dialysis
ter, Johns Hopkins Hospital, Baltimore, MD.
exhibited substantially improved survival for the
Received March 15, 1996; accepted in revised form May
14, 1996. first 2 years on dialysis, perhaps because they
Supported by a Clinical Center Grant from the National did not start dialysis with hypoalbuminemia. In
Institutes of Health, Bethesda, MD (RR-00722). the present study this treatment has been ex-
Address reprint requests to Mackenzie Walser, MD, Johns tended to additional patients with the nephrotic
Hopkins School of Medicine, 725 N Wolfe St, Baltimore, MD
2120.5.
syndrome. Those with less severe renal insuffi-
0 1996 by the National Kidney Foundation, Inc. ciency have exhibited remission or near-remis-
0272~6386/96/2803-0006$3.00/O sion of their nephrotic syndrome.

354 American Journal of Kidney Diseases, Vol 28, No 3 (September), 1996: pp 354-364
TREATMENT OF NEPHROSIS WITH A LOW-PROTEIN DIET 355

Table 1. Summary of Presenting Data in Patients Studied, Arranged in Order of Increasing Glomerular
Filtration Rate

GFR
(mu
Patient Age Hist SUN Cr TWII-IS Chol LDL HDL TG Alb minl3 uprotv
NO. Sex (yr) Diagnosis (mo) Ster? ACEI? (mg/dL) (mg/dL) (mg/dL) (mg/dL) (mg/dL) (mg/dL) (mg/dL) WV m? (g/d)

23 F 61 MN 60 No No 61 6.8 244 209 - - 114 3.3 7 5.8

44 M 31 IDDM ‘71 No Yes 41 5.3 162 201 142 48 62 2.6 9 5.5
121 F 45 FSGS’ 12 Yes No 21 2.6 132 461 t 342t 44t 376t 1.5 9 9.5
143 M 63 MN* 12 Yes No 60 5.7 204 306 237 49 101 2.9 10 11 .o
96 M 62 FSGS 6 Yes No 74 2.4 151 526 361 32 664 1.7 15 10.9
65 M 37 IDDM 4 NO YES 38 3.8 214 340 273 47 100 3.2 20 14.4
60 F 72 FGN’ 23 No No 41 2.7 la3 273 200 47 148 3.2 24 3.5
aa M 66 RAS, 1 No No 52 3.3 205 270t 1817 6at 106t 3.0 24 6.7
IDDM
120 M 64 AMYL’ 24 Yes No 16 1.1 115 583 474 33 379 1.3 24 9.9
71 F 39 MPGN’ 48 Yes No 31 2.2 185 4437 313t w 153t 2.3 28 23.8
46 F 78 MPGN* 96 Yes No 34 1.9 234 384 305 44 177 3.1 30 6.4
142 F 50 IDDM, 4 No No 39 1.2 ia9 411 210 49 a25 2.5 32 14.0
FSGS*
77 F 53 FSGS’ 14 No No 18 1.2 170 3401 242t 3gt 294t 2.3 41 10.7
100 M 62 IDDM” 5 No No 45 1.5 227 487 378 71 188 2.5 58 a.2
125 F 31 FSGS’ 14 No No 11 1.2 219 468 363 53 262 2.5 61 21.4
136 F 31 FSGS’ 4 No No 20 0.9 194 469 337 119 aa 2.7 69 5.4
Mean + SD 37 ? 27 20 i 25 38~172.7i1.8189i35386~110291~8656~23252~2112.5i0.629~1910.4~5.5

Abbreviations: FSGS, focal segmental glomerulosclerosis: IDDM, insulin-dependent diabetes mellitus; MN, membranous nephropathy; MPGN, membranoprolif-
erative glomerulonephritis; FGN, fibrillary glomerulonephritis; AMYL, amyloidosis; RAS, renal artery stenosis: Hist, duration of nephrotic syndrome prior to starting
diet; SW?, prior course of steroids; ACEI?, taking an angiotensin-converting enzyme inhibitor; SUN, serum urea nitrogen; 0, serum creatinine; Trans, serum
transferrin; Chol, serum cholesterol; LDL, serum low-density lipoprotein cholesterol; HDL, serum high-density lipoprotein cholesterol: TG, serum triglycerides:
Alb, serum albumin; GFR, glomerular filtration rate; UprotV, 24.hour urinary protein excretion.
*Biopsy.
tTaking an antihyperlipidemic drug

MATERIALS AND METHODS patient no. 142, simvastatin was withdrawn 4 weeks before
her first visit.
All 16 adult patients who presented with the nephrotic
The GFR before (and at intervals after) starting the diet
syndrome and who agreed to consume the diet described
was measured as the urinary clearance of 99mTc-DTPA, aver-
below are the subject of this report, including the five listed
aged over three collection periods, after a water load of 1
above, analyzed retrospectively, and 11 others studied pro-
L.30 Nondiabetics were fasted for this procedure, but diabetics
spectively. Nephrotic syndrome was defined for the purposes
were permitted a light breakfast. The GFR was expressed in
of this study as more than 3 g/d of proteinuria and hypoal-
relation to ht’ rather than surface area because weight, and
buminemia (<3.4 g/dL). Presenting features of these 16 pa-
therefore surface area, changes but height does not, at least
tients are summarized in Table 1, arranged in order of increas-
in the short term, and subjects who lose weight progressively
ing glomerular filtration rate (GFR), which varied from 7 to
may appear to be gaining renal function (and conversely)
69 mL/min/3 mz of height squared (ht’). Prior duration of the
when surface area is used as a referent.’ Average normal
nephrotic syndrome averaged 20 months.
The diet has been described in detail previously.30 It con- height is approximately & m (5 ft 8 in); hence, normal ht*
tains 0.3 g/kg ideal body weight of mixed-quality protein and is 3 m*. Measurements of GFR in which urine flow was
7 to 9 mg/kg ideal body weight of phosphorus. All patients judged too low were censored according to a procedure de-
took a supplement of essential amino acids or, in four pa- scribed previously8,3’; this caused eight of a total of 170 GFR
tients, alternating amino acid and ketoacid3’ supplements in measurements to be censored.
divided doses with meals; total daily dose was either 10 g The GFR values in many of these patients are lower than
(four patients) or 20 g (12 patients) of amino acids. In addi- their serum creatinine concentrations would lead one to ex-
tion, all patients took supplemental CaC03 sufficient to pro- pect. For example, five patients with serum creatinine concen-
vide 1 g elemental calcium, a multivitamin preparation (usu- trations within normal limits (nos. 120, 142, 77, 125, and
ally Nephrocaps; Fleming & Co, Fenton, MO), and 10 mgl 136) nevertheless had subnormal GFRs. In the nine patients
d of ZnS04. If serum iron was low, FeS04 was added. No with serum creatinine concentration values greater than 2.0
patients required erythropoietin injections. Two patients (nos. mg/dL, GFRs predicted from serum creatinine concentration,
44 and 65) continuously received an angiotensin-converting age, weight, sex, and height, using formulas derived from
enzyme inhibitor. NaHC03, additional antihypertensives, di- nonnephrotic patients that we have studied,32 were 24% -t
uretics, KCl, colchicine, allopurinol, insulin, and other drugs 20% (SEM; P = NS) higher than the observed GFRs. This
were prescribed as needed. Four patients (nos. 121, 88, 71, discrepancy has been noted previously in nephrotic patients
and 77) continued to receive antihyperlipidemic drugs; in and has been attributed to relatively high rates of tubular
356 WALSER, HILL, AND TOMALIS

Table 2. Changes in Nephrotic Indices During Dietary Treatment in Patients With Glomerular Filtration Rates
530 mUmin/ m2

Months No. of Patients Change in Change in Serum Change in Serum Change in

of Diet Remaining Proteinuria (g/d) Albumin (g/dL) Cholesterol (mg/dL) GFR (mUmin/ m*)

2 9 -0.9 ? 1.1 +0.2 +- 0.1 -50 +- 18* -2 2 1

3 8 -1.5 2 1.3 +0.3 2 0.1t -47 2 19-t -2 2 2
4 8 -2.9 t 1.4 +0.4 i 0.1* -69 t 21* -4 + 2
5 7 -3.9 2 1.4t +0.4 2 0.1* -85 ? 21$ -122

NOTE. Data are given as mean values IT SEM.

*/J < 0.025.
tf < 0.05.
*P < 0.01.

secretion of creatinine.33 Another possibility is that these pa-
tients have lower muscle mass.
All blood and urine chemical measurements were per-
formed in the Johns Hopkins Hospital Chemical Laboratory
or in the MarylandMedical Laboratory. Serum albumin was
determined photometrically in a Hitachi Analyzer (San Jose,
CA) using bromcresol green.
RESULTS
Effects of Dietary Treatment
The response to treatment of patients with
GFRs 530 mL/min/3 m2 of ht2 was strikingly
different from the response of the five patients
with higher GFRs. Consequently, the results will
be described separately for these two groups, des-
ignated group I and group II.
Group I patients’ responses to treatment are
presented in Table 2 as changes in serum levels
of albumin and cholesterol, proteinuria, and GFR
with duration of treatment. This analysis is poten-
tially misleading in that patients who drop out
and start dialysis bias the results toward a pro-
gressively less severe population remaining with
time. However, the results are scarcely different
when the patients who dropped out are excluded
from the analysis. Stated another way, the results
shown are relevant only to patients whose renal
failure is not so severe as to preclude their com-
pleting several months of dietary therapy. Mean
values and standard deviations of these four pa-
rameters at the start of treatment in all 11 patients
(Table 1) were as follows: proteinuria 9.8 + 5.4
g/d, serum albumin 2.6 + 0.7 g/dL, serum choles-
terol 363 ? 121 mg/dL, and GFR 18 t 8 mL/
min/3 m2 of ht’.
The results show that proteinuria, hypoalbumin-
emia, and hypercholesterolemia all improve pro-
gressively, on the average, with duration of di-
etary treatment, the most significant changes be-
ing in serum cholesterol, which decreased 85 _t
21 mg/dL by 5 months (P < 0.01). The largest
change proportionately was in proteinuria, which
decreased 37% k 14 by 5 months (P < 0.05).
Mean serum albumin concentration increased
significantly (P < 0.05) by 3 months and re-
mained 0.4 k 0.1 g/dL above the starting level
at 4 and 5 months. These patients were all treated
until dialysis became necessary; this interval var-
ied from 1 month (patients no. 120, 121, and
143) to 3 years (patient no. 71; mean, 9.7 2 10.5
[SD] months; median, 6 months), depending on
the severity of their renal insufficiency and their
nephrotic syndrome at the start, as well as the
rate of loss of renal function. Patients no. 71 and
46, with the longest intervals, were the patients
whose initial GFRs were the highest of those in
group I (28 and 30 mL/min, respectively).
Group II patients’ responses to treatment are
shown in Figs 1 to 4 and in Table 3. Proteinuria
decreased from 9.3 + 6.8 (SD) g/d to 1.9 _t 1.1
g/d (P < O.OOl), but the speed of this response
varied from 3 months (in patient no. 142) to 14
months (patient no. 77). In patient no. 77, a fam-
ily get-together at 13 months, during which the
patient went off the diet completely, resulted in
a transient increase in proteinuria (Fig 1) and
decrease in serum albumin (Fig 2).
Serum albumin increased rapidly to normal in
patients no. 100 and 142, and more slowly in the
others. In patient no. 136, an erroneous labora-
tory report of a very low hematocrit at 6 months
led to discontinuance of the diet and resumption
of a normal diet for 10 days; serum albumin
TREATMENT OF NEPHROSIS WITH A LOW-PROTEIN DIET 357

Table 3. Effect of Treatment on Nephrotic Syndrome in Five Patients With Glomerular Filtration Rates Greater
Than 30 mUmin/ m2

Serum Cholesterol
Proteinuria (g/d) Serum Albumin (g/dL) (w/W GFR (mUmin/ m’)

Patient No. Initial Final Initial Final Initial Final Initial Final

142 14.0 0.5 2.5 3.7 411 264 32 94

77 10.7 2.2 2.3 4.1 340 236 41 54
100 8.2 1.4 2.5 4.1 487 204 58 65
125 21.4 2.8 2.5 4.0 468 252 61 68
136 5.4 3.3 2.7 3.5 469 319 69 69
Mean 2 SD 9.3 2 6.8 1.Y 2 1.1 2.5 t 0.1 3.8* 2 0.3 435 2 54 255* -t 38 52 t 14 702 13

*/=J < 0.001.

decreased to a value (2.1 g/dL) lower than its
initial value (2.7 g/dL), but recovered rapidly and
was within the normal range by 10 months. There
was no associated change in proteinuria (Fig l),
but serum cholesterol, which had been decreas-
ing slowly, increased for a month or so (Fig 3).
In the other patients, serum cholesterol decreased
progressively, but did not become normal (ie,
<200 mg/dL) in any subject. In patient no. 142,
serum cholesterol increased at first, probably be-
cause simvastatin had been discontinued only a
few weeks previously. The GFR increased at
widely differing rates and to differing degrees
(Fig 4). The mean change in GFR at the end of
dietary therapy was +50% (P = NS; Table 3),
but, as shown in Fig 4, GFR clearly increased in
only three of the five patients and very little in
one of these three (no. 77). The GFR increased
threefold in patient no. 142.” The GFR did not
decrease during dietary therapy in any patient.
Long-term changes in GFR are shown below.
Average protein intake (excluding supplement
nitrogen, which amounted to 1 to 2 g/d, de-
pending on the dose of essential amino acids),
estimated from average urinary excretion rates
during treatment of urea nitrogen plus protein
nitrogen plus 0.031 g/kg,35 was 0.75 t 0.05 g/
kg in group I and 0.96 t 0.15 g/kg in group II
before starting the diet and 0.42 i: 0.04 g/kg in
group I and 0.50 t 0.04 g/kg in group II during
the treatment period. This level of compliance is
similar to that which we have observed in previ-
ous studies in nonnephrotic patients.30
Average values for mean arterial pressure in
groups I and II were 111 t 5 mm Hg and 91 i
1 mm Hg, respectively, before treatment; average
changes in mean arterial pressure during follow-
up were -8 t 2 mm Hg (P < 0.05) and +6 rt
5 mm Hg, respectively.

Long-Term Follow-up of Group II Patients

The effects of reintroducing normal foods
were examined in four of the five patients in

* This surprising increase in GFR was not attributable to

Fig 1. Sequential changes in proteinuria in five pa-
correction of dehydration or hypovolemia: mean arterial pres-
tients with GFRs greater than 30 mL/minR m’ of hi*
when placed on a supplemented very low-protein diet. sure at the first four visits (measured sitting) was 93, 93, 95,
Symbols shown in parentheses represent measure- and 94 mm Hg, respectively; weight was 84.7, 84.5, 83.6
ments made during intervals when the patients were and 85.4 kg, respectively. The urine flow rate was ~2.8 mL/
not on the diet. min.
 WALSER, HILL, AND TOMALIS

Fig 4. Sequential changes in GFR in the same pa-

tients as in Fig 1. The shaded area represents part of
the normal range for subjects aged 50 years.34 The
Fig 2. Sequential changes in serum albumin in the final GFR in patient no. 142 was obtained at the end
same patients as in Fig 1. The shaded area represents of the food reintroduction experiment shown in Fig 7.
part of the normal range.

creased. At a visit 3 months ago, the decision was

group II (in addition to the two inadvertent in-
stances shown in Figs 1 to 4). The fifth patient
(no. 125) has been lost to follow-up. In patient
no. 77, various foods were introduced one at a
time, starting 14 months after the diet was begun.
Specifically, these were egg white, casein (as Ca-
set; Mead Johnson, Evansville, IN), fat-free
cheese, lima beans, cheddar cheese, eggs, bread,
tuna fish, bacon, sausage, Mexican food, and tur-
key. The patient finally resumed a normal diet
and discontinued amino acids 20 months after
her first visit. As Fig 5 shows, none of these
measures has had any clearcut effect on protein-
uria or serum albumin. Serum cholesterol in-
creased 25 months after the start of treatment
when simvastatin was discontinued (because of
elevated transaminase levels). The GFR re-
mained constant until recently, when it has in-
made to resume the diet to determine whether
hypercholesterolemia would again respond; 2
months later serum cholesterol had decreased
from 419 mg/dL to 310 mg/dL (not shown).
In patient no. 100, individual foods were rein-
troduced successively after 5 months, including
casein, bread, eggs, cheese, pasta, pork, beef,
nuts, and chocolate. A normal diet was finally
resumed, without amino acid supplements. None
of these foods has had any definite effect on pro-
teinuria, serum cholesterol, or serum albumin

Fig 5. Effect on the nephrotic syndrome of reintro-

ducing normal foods one at a time in patient no. 77.
Apart from a modest increase in proteinuria, no
change in parameters of the nephrotic state is seen.
At the arrow, pravastatin was withdrawn because of
elevated serum transaminase levels; serum choles-
Fig 3. Sequential changes in serum cholesterol in terol increased. The GFR has recently increased. Time
the same patients as in Fig 1. is measured in months from the start of treatment.
TREATMENT OF NEPHROSIS WITH A LOW-PROTEIN DIET 359

0.7 g/d. However, serum cholesterol remains

high (259 mg/dL). Thus, she has been in remis-
sion (in terms of the definition of nephrotic syn-
drome used in this study) for 6 months, except
for a single value of proteinuria of 4.5 g/d at
Serum cholesterol, mgldl month 10.5.
150
Serum albumin, gfdl DISCUSSION
In group I patients, the decline in GFR is
doubtless one factor contributing to the decrease

Fig 6. Effect on the nephrotic syndrome of reintro-

r
120 GFR, ml/min

ducing normal foods successively in patient no. 100.

There was some increase in serum cholesterol, but
little change in proteinuria or in serum albumin con-
centration. The GFR has recently increased to normal.
Time is measured in months from the start of treat-
ment.

80

concentration (Fig 6). Finally, at month 12, the

patient resumed a normal diet except that he
3oo,cl
avoids beef, fish, “dark’ sodas, and dark choco-
late because he believes that these foods contrib-
ute to proteinuria. Serum albumin and cholesterol 200 L 1 I I I I I I J
have remained normal and proteinuria has re-
mained less than 2 g/d. Thus, he has been in
remission for 18 months.
In patient no. 142, beef (approximately 30 g
protein/d) was reintroduced after 3 months, for
a period of 2 weeks. Proteinuria did not change
appreciably, but serum cholesterol increased (Fig
7). Substitution of milk (one qt/d, equivalent to
2 Proteinuria, g/day
30 g protein) for the following 2 weeks had little
effect. Because of persistent hypercholesterol-
emia, the patient then resumed the diet, but uri-
nary urea nitrogen determinations show that her
compliance has been poor. Nevertheless, serum
cholesterol has remained only moderately ele-
vated, serum albumin has remained normal, and
0
proteinuria has progressively decreased, being 4 6 8 10
undetectable (except as microalbuminuria) at her Months
last visit. The GFR has remained normal or
nearly normal. Thus, she has been in remission Fig 7. Effect of reintroducing normal foods in pa-
for 7 months and has no objective signs of renal tient no. 142. First beef and then milk (at month 3.2)
were introduced. Beef led to an increase in serum cho-
disease. lesterol, but neither proteinuria nor serum albumin
Patient no. 136 has not fared as well, as shown changed appreciably. Thereafter, a normal or nearly
in Fig 8, until her last two visits, when, despite normal diet was consumed, but was supplemented by
essential amino acids. All signs of renal disease have
consuming a normal diet, she exhibited an in- disappeared. Time is measured in months from the
crease in GFR to normal and proteinuria of only start of treatment.
360 WALSER, HILL, AND TOMALIS

control group of a recent study had a spontaneous

remission of their nephrotic syndrome,37 and in
nephrotic syndrome due to focal segmental glo-
merulosclerosis, one of 88 patients was observed
to remit spontaneously in one study3’ and none
80
of eight patients in another.3g
Table 4 summarizes the results of all 14 previ-
ous studies that we have been able to identify
of protein-restricted diets in nephrotic subjects
60 : (excluding abstracts and our preliminary re-
port”). Most, but not all, of these previously re-
Serum cholesterol, mg/dl
350 ported patients presented with hypoalbuminemia
(<3.4 g/dL) and nephrotic-range proteinuria
(>3.5 g/d). Although some increase in serum
albumin and some decrease in proteinuria was
commonly observed, not one of these 202 pa-
250 L
tients exhibited normalization of serum albumin
Serum albumin, g/dl concentration and reduction of proteinuria to less
4.0 than 3.5 g/d in response to dietary treatment, in
contrast to all five of the group II patients in-
3.5 cluded in the present report. In most cases, GFRs
are not reported in these studies, but when they
3.0 t I I I I I I I are reported, GFRs decreased.
The first study of Barsotti et aI” is the only
one that used a very low-protein, supplemented
4r
diet, in only four patients (the other four patients
received 0.55 g/kg of protein). Of these four pa-
3- tients, only one (“DBS”) presented with hypo-
proteinemia, which was corrected by the diet.
2- Proteinuria decreased in all eight patients, to an
average of 3.1 g/d. Serum albumin increased, on
l- the average, by 0.5 g/dL. Thus, this study comes
Proteinuria, g/day closer, in design and in results, to the present
O- study than any other study reported. Two later
10 12 14 studies by the same investigators’4,‘5 used a less
Months restrictive diet (0.7 g/kg of protein).
In six of these 14 studies, some sort of random-
Fig 8. Effect of reintroducing normal foods in pa- ization was carried out. All six of these random-
tient no. 136. Hypercholesterolemia persists, but GFR ized studies were of short duration (2 to 16
has increased to normal. Time is measured in months
from the start of treatment. weeks). It should be pointed out that randomiza-
tion of half of the patients to no treatment (in-
cluding no steroids or immunosuppressive drugs)
in proteinuria and perhaps to the decline in serum for periods longer than 1 year (as in the present
cholesterol.36 In group II patients, all of whom study) is not feasible and may not be ethical.
had either diabetic nephropathy, focal segmental However, the point of the present study was not
glomerular sclerosis, or both, it is clear that the to determine how often remission is induced by
improvement seen was caused by the diet and diet (which would require a much larger series
not by chance or secondary to decreasing GFR, of patients with various disorders) but rather to
because GFR tended to increase. Furthermore, demonstrate that remission may be induced by
spontaneous remissions are rare in these disor- diet. Since spontaneous remission is rare in focal
ders: in type I diabetes, one of 66 patients in the segmental glomerulosclerosis and in diabetic ne-
TREATMENT OF NEPHROSIS WITH A LOW-PROTEIN DIET 361

Table 4. Summary of Published Studies of Protein-Restricted Diets in Nephrotic Patients

Serum Albumin
Diet Protein (g/kg) (g/W Proteinuria (g/d)
Duration
Ref Year Design* n 0-4 Cont Exptlt Cont Exptl Cont Exptl

9 1986 AEVBA 9 0.5 1.6 0.8 2.0 2.2 12.8 9.1

10 1988 AB 8 8-25 1.3 0.3*-0.55* 3.3 3.8 5.7 3.1
11 1988 AB 11 1 1.8 0.6 3.5 3.5 4.0 3.1
12 1988 ABIBA lO§ 0.75 1.2 0.7 3.9 3.9 3.9 2.4
13 1989 ABCIACB 12 1 1 .o, 2.0 0.5 2.7, 2.8 2.9 8.2, 9.2 6.1
14 1990 AB 13 4 1.0 0.7 2.6 2.9 8.7 5.6
15 1991 AB 20 2-12 i-i.3 0.7 2.6 2.9 7.8 5.2
16 1991 AB 12 0.75 1.6 0.8 3.4 3.3 6.9 6.5
17 1991 N/B 14/I 1 1.2 0.7 3.3 3.7 3.2 1.9
18 1991 AB 13 1.25 ? 0.6 3.3 3.3 4.1 3.5
19 1991 AEVBA 9 3 1.3 0.6 3.1 3.0 3.5 3.1
20, 1991, ABIBA 24 3 1.1 0.7 2.1 2.3 6.7 5.6
21 1992
22 1992 AB 20 2 ?1.1 0.7 “Unchanged” 5.9 4.0
23, 1993, ABCA/ 20 4 1.1 0.7 2.9 2.9 6.1 5.1
24 1993 ACBA
25 1995 AB 7 2 l-l.5 0.5-0.7 3.0 3.1 7.4 6.3
Group I AB 11 l-36 0.751 0.3 2.6 3.0 9.8 5.9
Group II AB 5 3-14 0.961 0.3 2.5 3.8 9.3 1.9

Abbreviations: Cont, control diet; Exptl, experimental diet.

*A, control diet; B, experimental diet; C, high-protein diet (2 g/kg); D, low-protein diet (0.7 g/kg) plus fish oil; /, randomiza-
tion to sequences given before and after /; N, patients “assigned” to control or experimental diet.
TPrescribed.
*Four of these eight patients were prescribed the very low-protein diet (0.3 g/kg) and four a low-protein diet (0.55 g/
kg); all eight received in addition 0.15 g/kg of a mixture of amino acids and keto-analogues.
$Only six of 10 patients had nephrotic-range proteinuria.
/ILess than half of these patients had nephrotic-range proteinuria.
IEstimated from weight, 24-hour urinary urea nitrogen, and proteinuria.35

phropathy, it is clear that the remissions that we
have observed in five consecutive patients with
these diagnoses and GFRs greater than 30 mL/
min cannot be due to chance. Further study will
be required to establish how regularly this re-
sponse is demonstrable.
The difference between the present results and
these previous reports suggests that some compo-
nent of the diet, not eliminated in these previous
studies, that was totally eliminated in the present
study, may have been responsible for the failure
of these previously studied patients to exhibit
remissions. Several high-protein foods were
completely eliminated in our patients, including
meat, fish, poultry, cheese, and milk.
This raises the possibility of food allergy,
which has previously been suggested to be a
causative factor in the nephrotic syndrome40-43
based on incomplete evidence. Our observation
that four patients were later able to tolerate nor-
mal foods without exacerbation of their nephrotic
syndrome argues against food allergy. Further-
more, it is difficult to conceive that either diabetic
nephropathy or focal segmental glomerulosclero-
sis can be forms of food allergy.
There is additional evidence that specific foods
may play a role in the nephrotic syndrome. For
example, chicken and fish are reported to cause
less hyperfiltration in diabetic subjects than
meat.& The results of D’Amico and Gentile,22
Gentile et aLz3 and Barsotti and colleagues14,‘5
are also consistent with the possibility that re-
placement of animal protein in the diet with vege-
table protein ameliorates the nephrotic syndrome.
Another possible mechanism by which protein
restriction might ameliorate the nephrotic syn-
drome is suggested by a preliminary report by
Lerma et a1,45 who found that glomerular cells

from doxornbicin-treated rats fed a low-protein
diet did not exhibit increased production of tumor
necrosis factor and platelet-activating factor,
while rats fed a normal protein diet developed
much more severe nephrosis and their glomerular
cells produced increased quantities of these me-
diators.
The essential amino acid supplement may have
exerted some effect in these patients other than
providing sufficient amino acids to prevent pro-
tein deficiency on this diet. For example, it is
conceivable that the rate of oxidation of one or
more essential amino acids is abnormally high
in nephrotic patients, and the provision of addi-
tional quantities of the lacking amino acid(s) may
have corrected this deficiency. Some support for
this idea can be found in the observation that
plasma levels of branched-chain ketoacids are
subnormal in nephrotic patients, even though
plasma levels of branched-chain amino acids are
norma1.46 Determination of the levels of circulat-
ing amino acids and/or ketoacids before and after
treatment might provide clues to confirm this hy-
pothesis.
On the other hand, only certain dietary amino
acids may be responsible for the renal injury ap-
parently associated with the ingestion of a normal
diet in these subjects prior to treatment. Kaysen
and colleagues have shown that neither arginine
plus proline plus glutamate47 nor branched-chain
amino acids4* added selectively to the diet of
nephrotic rats, aggravate albuminuria or hypoal-
buminemia. Thus, if certain dietary amino acids
are responsible, arginine, proline, glutamate, and
branched-chain amino acids are unlikely candi-
dates. On the other hand, according to el-Gayar
et a1>9 plasma glutamate and hydroxyproline are
both elevated in nephrotic patients, suggesting
that these nonessential amino acids may play a
role.
Other components of proteinaceous food could
be responsible, such as phosphoproteins, nucleic
acids, inorganic phosphate, or specific lipids.
Further studies will be required to identify which
components of the normal diet, when removed,
were responsible for the improvement that oc-
curred in these patients.
The absence of a relapse in three patients who
resumed a normal diet after remission, when con-
sidered in conjunction with the apparent worsen-
ing in two patients who went off the diet at an
WALSER, HILL, AND TOMALIS
early stage, suggests that the susceptibility of the
kidney to diet-induced damage was a transient
phenomenon. If correct, this finding will make
study of the dietary factors responsible more dif-
ficult.
The increase in GFR seen in four of the five
patients with less severe renal insufficiency is of
particular interest. We have rarely observed an
increase in GFR in nonnephrotic patients during
dietary treatment.30 The increase in GFR is prob-
ably a consequence of the correction of hypopro-
teinemia. Manning” has shown that dogs sub-
jected to repeated plasmapheresis over a 34-day
period, until their plasma protein concentration
is reduced 65%, show a 25% reduction in GFR,
in contrast to the results of acute experiments.
Although we did not perform second biopsies
in these patients, it seems likely that the abnor-
malities present on their first biopsies are at least
partially reversed. These abnormalities may be
in part a consequence of proteinuria, rather than
a cause.6’7 According to Ting et a1,51the anatomic
causes of reduced GFR in patients with membra-
nous nephropathy are (1) a pronounced reduction
in filtration slit frequency owing to epithelial
podocyte broadening and (2) marked thickening
of the glomerular basement membrane. These are
potentially reversible causes of hypofiltration.
What is clear from this work is that the com-
mon dictum that nephrotic subjects should not
be given severely protein-restricted diets needs
to be reconsidered. Not only did only one of
these 16 patients exhibit worsening of hypoal-
buminemia on this very restricted diet, but a
number of them showed substantial long-term
improvement. In contrast with other therapies for
the nephrotic syndrome, such as steroids or im-
munosuppressive drugs, this form of treatment,
although perhaps more arduous for the patient,
is safe and sometimes surprisingly effective. It
can be argued, therefore, that a course of dietary
therapy should be tried before attempts to treat
the nephrotic syndrome in adults with drugs. In
children, however, the protein requirement for
growth may limit the effectiveness of this form
of treatment.
ACKNOWLEDGMENT
The authors are grateful to Dr Luis F. Gimenez for referring
many of these patients.
TREATMENT OF NEPHROSIS WITH A LOW-PROTEIN DIET 363

REFERENCES Gentile MG, Remuzzi G: Low-protein diet and glomerular

size-selective function in membranous glomerulopathy. Am
1. Lewis EJ: Management of nephrotic syndrome in J Kidney Dis 17:317-322, 1991
adults. Kidney Dis 8:461-521, 1988 20. D’Amico G, Remuzzi G, Maschio G, Gentile MG,
2. Epstein AA: The nature of treatment of chronic paren- Gotti E, Oldrizzi L, Manna G, Mecca G, Rugiu C, Fellin G:
chymal nephritis (nephrosis). JAMA 69:444-449, 1917 Effect of dietary protein and lipids in patients with membra-
3. Blainey JD: High protein diets in the treatment of ne- nous nephropathy and nephrotic syndrome. Clin Nephrol
phrotic syndrome. Clin Sci 13:567-581, 1954 351237-242, 1991
4. Farr LE: The assimilation of protein by young children 21. D’Amico G, Gentile MG, Manna G, Fellin G, Ciceri
with the nephrotic syndrome. Am J Med Sci 195:70-76, 1938 R, Cofano F, Petrini C? Lavarda F, Peroline S, Porrini M:
5. Keutmann EH, Bassett SH: Dietary protein in hemor- Effect of vegetarian soy diet on hyperlipidemia in nephrotic
rhagic Bright’s disease. II. The effect of diet on serum pro- syndrome. Lancet 339:113-1134, 1992
teins, proteinuria and tissue proteins. J Clin Invest 14:871- 22. D’ Amico G, Gentile MG: Effect of dietary manipula-
888, 1935 tion on the lipid abnormalities and urinary protein loss in
6. Brenner BM, Meyer TW, Hostetter TH: Dietary protein nephrotic patients. Miner Electrolyte Metab 18:203-206,
intake and the progressive nature of kidney disease. N Engl 1992
J Med 307:652-659, 1982 23. Gentile MG, Fellin G, Cofano F, Delle Fave A, Manna
7. Thomas ME, Schreiner GF: Contribution of proteinuria G, Ciceri R, Petrini C, Lavarda F, Pozzi F, D’Amico G:
to progressive renal injury-Consequences of tubular uptake Treatment of proteinuric patients with a vegetarian soy diet
of fatty acid bearing albumin. Am J Nephrol 13:385-398, and fish oil. Clin Nephrol 40:315-320, 1993
1993 24. D’ Amico G; Gentile MG: Influence of diet on lipid
8. Walser M: Progression of chronic renal failure in man. abnormalities in human renal disease. Am J Kidney Dis
Kidney Int 37:1195-1210, 1990 22:151-157: 1993
9. Kaysen GA, Gambertoglio J, Jimenez I, Jones H, 25. Gansevoort RT, de Zeeuw D, de Jong PE: Additive
Hutchison FN: Effect of dietary protein intake on albumin antiproteinuric effect of ACE inhibition and a low-protein
homeostasis in nephrotic patients. Kidney Int 29:572-577, diet in human renal disease. Nephrol Dial Transplant 10:497-
1986 504, 1995
10. Barsotti G, Ciardella F, Morelli E, Cupisti A, Manto- 26. Feehally J, Baker F, Walls J: Dietary protein manipula-
vanelli A, Giovannetti S: Nutritional treatment of renal failure tion in experimental nephrotic syndrome. Nephron 50:247-
in type 1 diabetic nephropathy. Clin Nephrol 29:280-287, 252, 1988
1988 27. Fujisawa K, Yagasaki K, Adachi I, Funabiki R: Quali-
11. Wetzels JFM, Hoitsma AJ, Berden JHM, Koene RAP: tative and quantitative effects of dietary proteins on hyperlip-
Renal hemodynamic effects of a short-term high and low idemia and proteinuria in nephritic rats. Nutr Res 13:1055-
protein diet in patients with renal disease. Clin Nephrol 1066, 1994
30:42-47, 1988 28. Walser M: Does prolonged protein restriction preced-
12. Bending JJ, Dobbs RA, Keen H, Viberti GC: Renal ing dialysis lead to protein malnutrition at the onset of dial-
response to restricted protein intake in diabetic nephropathy. ysis? Kidney Int 44:1139-1144, 1993
Diabetes 37:1641-1646, 1988 29. Coresh J, Walser M, Hill S: Survival on dialysis
13. Mansy H, Goodship TH, Tapson JS, Hartkey GH, among chronic renal failure patients treated with a supple-
Keavey P, Wilkinson R: Effect of a high protein diet in mented low-protein diet before dialysis. J Am Sot Nephrol
patients with the nephrotic syndrome. Clin Sci 77:445-451, 6:1379-1385, 1995
1989 30. Walser M, Hill S, Ward L: Magder L: A crossover
14. Barsotti G, Cupisti A, Morelli E, Ciardella F, Giovan- comparison of progression of chronic renal failure: Ketoacids
netti S: Vegan supplemented diet in nephrotic syndrome. vs. amino acids. Kidney Int 43:933-939, 1993
Nephrol Dial Transplant 5:75-77, 1990 (suppl 1) 3 1. Walser M: Progression of renal failure, in Davison AM
15. Barsotti G, Morelli E, Cupisti A, Bertoncini P, Giovan- (ed): Nephrology, vol II. Proceedings of the Xth International
netti S: A special, supplemented ‘vegan’ diet for nephrotic Congress in Nephrology. East Sussex, UK, Holt-Saunders,
patients. Am J Nephrol 11:380-385, 1991 1988, pp 1155-1181
16. Don BR, Kaysen GA, Hut&son FN, Schambelan M: 32. Walser M, Drew HH, Guldan JL: Prediction of glo-
The effect of angiotensin-converting enzyme inhibition and merular filtration rate from serum creatinine concentration in
dietary protein restriction in the treatment of proteinuria. Am advanced chronic renal failure. Kidney Int 44:1145-1148,
J Kidney Dis 17:10-17, 1991 1993
17. Sugimoto T, Kikkawa HH, Shigeta Y: Effect of dietary 33. Carrie BJ, Golbetz HVC, Michaels AS, Myers BD:
protein restriction on proteinuria in non-insulin-dependent Creatinine: An inadequate filtration marker in glomerular dis-
diabetic patients with nephropathy. J Nutr Sci Vitaminol eases. Am J Med 69:117-182, 1980
37:S87-S92, 1991 34. Rolin HA, Hall PH: Age/sex stratified glomerular fil-
18. Shichiri M, Nishio Y, Ogura M, Marnmo F: Effect of tration rate (GFR) values defined by ‘251-iothalamate clear-
a low-protein, very-low-phosphorus diet on diabetic renal ance. J Am Sot Nephrol 3:347, 1992 (abstr)
insufficiency with proteinuria. Am J Kidney Dis l&26-32, 35. Maroni BJ, Mitch WE: A method for estimating nitro-
1991 gen intake of patients with chronic renal failure. Kidney Int
19. Remuzzi A, Perticicci E, Battaglia G, D’ Amico G, 2758-65, 1985
364
36. Moorhead JF, Persaud W, Verghese Z, Sweny P: Se-
mm cholesterol falls spontaneously in nephrotic patients with
progressive disease. Renal Fail 35:389-393, 1993
37. Hebert LA, Bain RP, Verme D, Cattran D, Whittier
FC, Tolchin F, Rohde RD, Lewis EJ, for the Collaborative
Study Group: Remission of nephrotic range proteinuria in
type 1 diabetes. Kidney Int 46:1688-1693, 1994
38. Rydel JJ, Korbet SM, Borok RZ, Schwartz MM: Focal
segmental glomerular sclerosis in adults: Presentation,
course, and response to treatment. Am J Kidney Dis 25:534-
542 1995
39. Nagai R, Cattran DC, Pei Y: Steroid therapy and prog-
nosis of focal segmental glomerulosclerosis in the elderly.
Clin Nephrol 42:18-21, 1994
40. Ferri C, Puccini R, Longombardo S, Paleologo G,
Migliorini P, Moriconi L, Pasero G, Cioni L: Low-antigen-
content diet in the treatment of patients with IgA nephropa-
thy. Nephrol Dial Transplant 8:1193-1198, 1993
41. Sieniawaska M, Szymanik-Grzelak H, Kowalewsla M,
Wasik M, Koleska D: The role of cow’s milk protein intoler-
ance in steroid-resistant nephrotic syndrome. Acta Pediatr
81:1007-1012, 1992
42. Laurent J, Lagrue G: Dietary manipulation for idio-
pathic nephrotic syndrome. Allergy 44:599-603, 1989
43. Sandberg DH, McIntosh GM, Bernstein CW, Carr R,
Strauss J: Severe steroid responsive nephrosis associated with
hypersensitivity. Lancet 1:388-401, 1977
WALSER, HILL, AND TOMALIS
44. Pecis M, de Azevedo MJ, Gross JL: Chicken and fish
diet reduces glomerular hyperfiltration in IDDM patients. Di-
abetes Care 17:665-672, 1994
45. Lerma JL, G6mez-Chiarri M, Gonzalez E, Ortiz A,
Hemando L, Egido J: Hypoprotein diet decrease proteinuria
and glomerular production of tumoral necrosis factor (TNF)
and of platelet activating factor (PAF) in experimental ne-
phrotic syndrome. Kidney Int 40:376, 1991 (abstr)
46. Cup&i A, Langer K, Barsotti G: Plasma levels of
branched chain keto acids in nephrotic patients. Nephron
63~237, 1993
47. Kaysen GA, Martin VI, Jones H Jr: Arginine augments
neither albuminuria nor albumin synthesis caused by high-
protein diets in nephrosis. Am J Physiol263:F907-F914, 1992
48. Kaysen GA, Al-Bander H, Martin VI, Jones H Jr,
Hutchison FN: Branched-chain amino acids augment neither
albuminuria nor albumin synthesis in nephrotic rats. Am J
Physiol260:F177-F184, 1991
49. el-Gayar A, Sobh M, elKholy A, Sallam S, Wafa E:
Alterations of plasma free amino acids in nephrotic syn-
drome. Int Urol Nephrol 26:707-712, 1994
50. Manning RD Jr:Effects of hypoproteinemia on renal
hemodynamics, arterial pressure, and fluid volume. Am J
Physiol 252:F91-98, 1987
51. Ting RH, Kristal B, Myers BD: The biophysical basis
of hypofiltration in nephrotic humans with membranous ne-
phropathy. Kidney Int 45:390-397, 1994