Renal Proximal Tubular Dysfunction and Paroxysmal Nocturnal Hemoglobinuria
ARTHUR L. RILEY, M.D. LAWRENCE M. RYAN, M.D. DONALD A. ROTH, M.D. Milwaukee, Wisconsin
A patient with paroxysmal nocturnal hemoglobinuria, who required many blood transfusions for hemolytic episodes, had a perslstent hyperchloremic metabolic acidosis. Bicarbonate Infusion demonstrated a large fractional excretion of bicarbonate (28.8 per cent at a plasma bicarbonate level of 23 megAlter) whidr was consistent with proximal renal tubular acidosis. Generalized aminoaciduria and decreased tubular reabsorption of phosphate were also present. Marked deposition of iron in renal proximal tubules was associated with these functional abnormalities. We believe that, as systemic acidosis can promote hemolysis in patients with paroxysmal nocturnal hemogloblnuria, hemolysis can lead, by way of iron deposltion in renal tubules, to further acidosis. This cycle should be interrupted with appropriate doses of bicarbonate. Paroxysmal nocturnal hemoglobinuria is a rare disease characterized by the onset of hemolytic episodes in adults which are frequently worse at night and during intercurrent infections [ 11. Associated with this is a significant tendency to spontaneous venous thrombosis which may involve multiple organ systems but most frequently and dramatically involves abdominal viscera. We report a case of paroxysmal nocturnal hemoglobinuria complicated by renal tubular acidosis which most likely was caused by the paroxysmal nocturnal hemoglobinuria and, if untreated, potentially exacerbates the primary disorder. disease
CASE REPORT
From the Renal Section, Medical Service, Veterans Administration Center, Wood (Milwaukee), Wisconsin; and the Department of Medicine, The Medical College of Wisconsin, Milwaukee 53226. This study was supported in part by NIH Grant No. 5 MO 1 RR00058. Requests for reprints should be addressed to Dr. Arthur L. Riley, Renal Section/ 111, Veterans Administration Center, Wood, Wisconsin 53193. Manuscript accepted January 30, 1976.
This 31 year old white man first sought medical attention in 1966, with complaints of episodes of abdominal pain and vomiting for one year, associated with intermittent dark urine and fever. Laboratory data available from another hospital showed that the blood urea nitrogen was 12 mg/iOO ml, hematocrit value 26 per cent and reticulocyte count 19 per cent; 2+ proteinuria was present: the serum lactic dehydrogenase and indirect bilirubin levels were elevated. An x-ray series of the upper gastrointestinal tract revealed thickened jejunal folds which prompted an exploratory laparotomy. At surgery there were no discernible abnormalities of the liver, spleen or kidneys. A small, superficial jejunal ulcer was discovered and resected. After an uneventful, postoperative course the patient was discharged, with a diagnosis of fever of undetermined origin, jejunal ulcer and “hemolytic reaction.”
January 1977
The American Journal of Medicine
Volume 62
125
RENAL PROXIMAL TUBULAR DYSFUNCTION
I
TABLE
Time (min)
Effect of Bicarbonate Paroxysmal Nocturnal
Urine Volume (ml/mitt)
0
Urine pH 8.0
AND PAROXYSMAL NOCTURNAL HEMOGLOBINURIA-RILEY
Infusion on Per Cent Excretion of Filtered Bicarbonate Hemoglobinuria and Hyperchloremic Acidosis
Venous PH
7.27
5 per cent sodium bicarbonate 30 60 90 120
6.0 6.3 5.0 6.3
7.5 7.5 8.0 7.5
“Calculated from determination anaerobically.
ET AL.
7.35 7.37 7.38 7.37
Urine Bicarbonate* (meq/liter) 6.8
at 100 ml/hour 10.2 20.8 29.3 27.9
Plasma Bicarbonate* (meq/liter)
Urine Plasma Creatinine Creatinine ~mg/lOO ml) (mg/lOO ml)
19.1
(= 60 meq/hour) 20.4 23.2 22.4 23.4
of pH and of carbon dioxide pressure and total carbon dioxide
After discharge the patient continued to be anemic and to have episodes of abdominal pain. He was referred to the Mayo Clinic where the diagnosis of paroxysmal nocturnal hemoglobinuria was established by means of positive sugar-water and Ham tests. Subsequent treatment consisted of the administration of prednisone, 40 to 60 mg daily, and a transfusion of 2 to 4 lJ of buffy-coat-poor packed red cells (washed in saline solution) approximately every four months. The patient was first seen at the Veterans Administration Center, Wood, Wisconsin, in March 1971, because of ab-
in Patient
with
% Excretion of Filtered Bicarbonate
20
2.4
4.3
15 10 20 10
2.4 2.4 2.4 2.4
8.0 21.5 15.7 28.6
intravenously
by microgasometer
on specimens collected
dominal pain and hemolytic episodes. At that time, his blood urea nitrogen was 18 mg/lOO ml, sodium 140 meq/liter, potassium 5.7 meq/liter, chloride 108 meq/liter, carbon dioxide 20 meq/liter and hematocrit value 20 per cent. Urinalysis revealed a pH of 5.0, specific gravity of 1.014 and 1-F protein. At this time, striking hemosiderinuria was present in association with evidence of iron deficiency (low serum iron, high total iron-binding capacity, hypochromic microcytosis and absent marrow iron stores). By November 197 1, the serum creatinine was 1.2 mg/lOO ml and creatinine clearance 130 liters/day. A 24-hour urine specimen yielded
Figure 1. Left, photomicrograph of iron stain of kidney biopsy specimen, principally reveals marked iron deposition in tubules. Right, electron micrograph. Proximal tubular cell is identified by brush border; large amounts of electron-dense material are present within the cytoplasm. These deposits were not seen in cytoplasm of distal tubular cells. Magnification X 160 (left) and X 2,500 (right).
126
January 1977
The American Journal of Medicine
Volume 62
RENAL PROXIMAL
TUBULAR DYSFUNCTION
3 g of protein; serum bicarbonate was 13 meq/liter with a urine pH of 5.5. Oral bicarbonate, in doses of 6 to 8 g/day, was required to maintain a serum carbon dioxide level of 14 meq/liter or higher, and the urine pf-l was then consistently higher than 6.0. The patient received 82 U of buffy-coat-poor packed red blood cells (washed in saline solution) between March 1971 and February 1974, when he was readmitted for evaluation of renal dysfunction. No history of exposure to nephrotoxins could be elicited: Serum creatinine was now 2.3 mg/lOO ml and creatinine clearance 55 liters/day. Parathyroid hormone level was 74 pleq/ml (measured in the laboratory of Dr. Claude Arnaud, Mayo Clinic) with serum calcium 9.0 mg/lOO ml, serum phosphorous 4.4 mg/lOO ml and serum magnesium 2.14 mg/lOO ml. Serum protein electrophoresis was within normal limits. The plasma free-hemoglobin level was 423 mg/lOO ml and plasma haptoglobin level 120 mg/lOO ml. The urine specimen was strongly positive for occult blood. Although a 24-hour urine collection for the determination of protein (Biuret method) revealed 21 g, only 3 per cent of this was albumin. Electrophoresis demonstrated that the majority (61 per cent) was in the alpha2 zone with the rest in the beta (17 per cent) and gamma (17 per cent) zones; these three zones stained strongly positive on hemoglobin stain. Generalized aminoaciduria was identified; no significant glycosuria could be detected. Tubular reabsorption of phosphate was 50 per cent of the filtered load when the patient was adhering to a general diet. The patient was receiving 100 meq of sodium bicarbonate orally per day when a bicarbonate loading test was performed by infusion of 5 per cent sodium bicarbonate solution at 100 ml/hour (Table I). Before the infusion, the plasma carbon dioxide was 20 meq/liter, and 4.6 per cent of the filtered load of bicarbonate was excreted. When the serum carbon dioxide level reached 23.6 rneq/liter during the infusion, 28 per cent of the filtered load was excreted. Liver and skin biopsy specimens were obtained to evaluate hepatomegaly and hyperpigmentation. Hemosiderosis was identified in both tissues. An intravenous pyelogram showed radiopaque kidneys before injection of contrast media. Both kidneys were enlarged and measured 17 cm in length; the cortical outline was smooth and the calyceal system appeared normal. Renal biopsy tissue was grossly rust colored and, on microscopic examination, showed marked proximal tubular siderosis as well as mononuclear interstitial infiltrates and minimal proliferative changes in some glomeruli (Figure 1). lmmunofluorescence was negative for immunoglobulins G, A, M, complement and fibrin. Electron microscopy revealed marked deposition of electron-dense particles within the cytoplasm of proximal tubular cells (Figure 1). No such deposition was present in the distal tubular cells. Subsequently, the patient has had two episodes of abdominal pain during which radiopaque findings were consistent with venous infarction of the small bowel. On both admissions, with the patient taking supplemental bicarbonate, the urine pH ranged between 7.0 and 8.0. At present, the patient is maintained on a regimen of prednisone, 5 mg daily; folic acid, 1 mg daily: and 4 to 6 U of packed cells at intervals of two to four months. To maintain a plasma bicarbonate level
AND PAROXYSMAL
10
14
NOCTURNAL
18
SERUM
22
HEMOGLOBINURIA-RILEY
26
CARBON
30
3a
DIOXIDE
38
ET AL.
(meq/liter)
42
Fimre 2. Bicarbonate reabsorption and excretion in our patient (0) during the intravenous infusion of 3 per cent sodium bicarbonate. The normal response to bicarbonate infusion, taken from Pitts et al. [ 31, is shown for comparison. GFR = glomerular filtration rate. in the low normal range, therapy with 120 meq/day of sodium bicarbonate is required. At no time has this patient demonstrated a nocturnal pattern of hemoglobinuria. COMMENTS
The patient had ample evidence of renal tubular acidosis, as demonstrated by a persistently alkaline urine in the presence of a metabolic hyperchloremic acidosis, but no evidence of a urinary tract infection. The identification of the acidosis as a proximal tubular defect was made by the abnormally low renal threshold for bicarbonate, with 4 per cent excretion of the filtered load at a serum concentration of 20 meq/liter and more than 20 per cent excretion at serum levels of 24 meq/liter [2]. The patient also required large amounts of bicarbonate supplement to increase the serum carbon dioxide to 20 meq/liter. In Figure 2, the responses to bicarbonate infusion in our patient are compared with the normal response to bicarbonate infusion [3]. The results suggest not only an early threshold for bicarbonate excretion but also a reduced capacity for bicarbonate reabsorption. Other proximal defects identified included aminoaciduria and decreased tubular reabsorption of phosphate (50 per cent during the adherence to a general diet). Recently, Muldowney et al. [4] demonstrated a reduction in maximal tubular reabsorptive capacity for bicarbonate in 15 patients with chronic renal failure of various etiologies with creatinine clearance ranging from 2.7 to 12.0 ml/min; most of these patients suffered either from pyelonephritis or from polycystic disease and had been uremic for relatively long periods. The
January 1977
The American Journal of Medicine
Volume 62
127
RENAL
PROXIMAL
TUBULAR
DYSFUNCTION
AND
PAROXYSMAL
NOCTURNAL
defects seemed related to the duration and degree of secondary hyperparathyroidism. Aminoaciduria was absent. It is unlikely that the proximal tubular defects in our patient can be explained only by “parathyroid acidosis” because (1) serum bicarbonate values were decreased and serum chloride values were increased in 1971 before the development of azotemia (creatinine clearance was 130 liters/day); (2) in 1974, the parathyroid hormone level was predictably high, but the mild to moderate azotemia at the time of bicarbonate infusion had been present less than one year; (3) the presence of aminoaciduria in our patient favors a more generalized proximal tubular disorder; and (4) the presence of marked iron deposition in the proximal tubules suggests that the defect in proximal tubular function is related to this morphologic lesion. Heavy metals are known to result in kidney damage, with manifestations which can include acute oliguric renal failure, chronic renal failure, the nephrotic syndrome and, importantly, abnormalities in proximal tubular function manifested by renal aminoaciduria, renal glycosuria and excessive loss of phosphate. Hyperchloremic acidosis, without further definition, is also described. Pathologic changes in epithelial cells of the proximal tubule correlate with this functional impairment
ia. A prominent part of our patient’s disease has been hemosiderosis as confirmed by liver, skin and renal biopsy. Whereas deposition of iron in the proximal tubules (and in the urine sediment) of patients with paroxysmal nocturnal hemoglobinuria is present even in the face of severe iron deficiency [6], the significant hemosiderosis in other tissues in this patient is the result of multiple transfusions. The deposition of iron pigment in the kidney was preferential for the proximal tubules, providing an anatomic correlate to his physiologic defect. Iron deposition in proximal tubules in various states HEMOLYSI / SYSTEMIC AFIDOSIS
BICARBONATE LOSS IN URINE
S t IRON DEPOSITION RENAL TUBULES (PR~XIMAL~
IN I
PROXIMAL
I
WRTA Vgure 3. Proposed scheme of interrelationship between paroxysmal nocturnal hemoglobinuria and renal tubular acidosis. The administration of appropriately large doses of bicarbonate can interrupt this cycle.
120
January 1977
The American Journal of Medicine
HEMOGLOBINURIA-RILEY
ET AL.
of intravascular hemolysis has been well described in the past [7-lo], but evidence of a concomitant functional defect has been lacking. In only one recent preliminary communication were specific defects of tubular function in patients with paroxysmal nocturnal hemoglobinurla considered; hyposthenuria was found in nine and abnormal phenolsulfonphthalein clearance in seven of the nine patients tested [ 111. Azotemia has occasionally been seen in patients with paroxysmal nocturnal hemoglobinuria, and chronic pyelonephritis [ 121, nephrosclerosis [ 131 or transfusion reactions [ 141 have been implicated. The significance of renal tubular acidosis in patients with paroxysmal nocturnal hemoglobinuria lies in the effect of systemic acidosis upon hemolysis. Strubing [ 151, in the first clinical description of paroxysmal nocturnal hemoglobinuria, noted the striking relationship of hemoglobinuria to sleep, attributing this to an accumulation of carbon dioxide or lactic acid. The Ham test (acid hemolysin test) is an in vitro determination, nearly specific for paroxysmal nocturnal hemoglobinuria, in which a pH of 6.5 augments hemolysis [ 161. Less dramatic alterations of pH, from 7.4 to 7.2 in vitro, also result in increased hemolysis [ 171. Local acidosis induced by arm exercise has been show to increase local hemolysis in a patient with paroxysmal nocturnal hemoglobinuria, associated with a decrease in antecubital vein pH from 7.48 to 7.33 [ 181. The administration of acetazolamide has had a similar effect [ 191. Finally, a large single dose of ammonium chloride has been shown to increase hemoglobinemia and hemoglobinuria with a simultaneous decrease of the arterial pH to 7.25 [20]. Therefore, it is reasonable to conclude that the lowering of blood pH in vivo can exacerbate paroxysmal nocturnal hemoglobinuria. One hypothesized mechanism is that acidosis enhances the sensitivity of paroxysmal nocturnal hemoglobinuria cells to complement. Our patient presented with paroxysmal nocturnal hemoglobinuria and coexistent renal tubular acidosis, aminoaciduria, proteinuria and decreased glomerular filtration. The biopsy finding of iron deposition, largely in the proximal convoluted tubules, correlates well with the functional tubular defects. In our review of the literature on paroxysmal nocturnal hemoglobinuria, no description of renal tubular acidosis was found. However, since the urine of patients with proximal renal tubular disease can be acidified when serum bicarbonate levels are very low, it is possible that the renal origin of such acidosis could be overlooked unless specifically pursued. The etiology of the tubular defect may well be related to the particular pattern of iron deposition seen, just as deposition of other metals, such as lead, uranium, cadmium and copper [21], may cause a spectrum of proximal tubular dysfunctions.
Volume 62
RENAL PROXIMAL TUBULAR DYSFUNCTION AND PAROXYSMAL NOCTURNAL HEMOGLOBINURIA-RILEY ET AL.
The therapeutic implications are important in that, as paroxysmal nocturnal hemoglobinuria may induce tubular changes, the resulting acidosis can accelerate hemolysis and lead to a self-perpetuating cycle of hemolysis causing proximal tubular dysfunction which increases acidosis and further hemolysis (Figure 3). Such a pathway can be interrupted by administering bicarbonate in large enough quantities to correct the metabolic acidosis. Patients with paroxysmal nocturnal
Dacie JV, Lewis SM: Paroxysmal nocturnal haemoglobinuria: clinical manifestations, haematology, and nature of the disease. Ser Haematol 3: 3, 1972. 2. Morris RC: Renal tubular acidosis: mechanisms, classification and implications. N Engl J Med 281: 1405, 1969. 3. Pitts RF, Ayer JL, Schiess WA: The renal regulation of acidbase balance in man. III. The reabsorption and excretion of bicarbonate. J Clin Invest 28: 35, 1949. 4. Muldowney FP, Donohoe JF, Carroll DV, et al.: Parathyroid acidosis in uraemia. Q J Med 41: 321. 1972. 5. Emerson BT: Metals and the kidney, chap 23. Renal Disease (Black DAK, ed), Philadelphia, F. A. Davis Co., 1967. 6. Hutt AT, Roger JF, Neustein HB: Renal pathology in paroxysmal nocturnal hemoglobinuria. Am J Med 31: 736, 1961. 7. Lamarca JT, Terroba JY, Diaz-Flores L: Hemoglobinuria paroxistica noctuma. II. lnsuficiencia Renal en el Curso de la Misma. Rev Clin Esp 28: 287, 1973. 8. Sigler AT, Furman EN, Zinkham WH, et al.: Severe intravascular hemolysis following surgical repair of endocardial cushion defects. Am J Med 35: 467, 1963. 9. Slater SD, Rahman M, Lindsay RM: Renal function in chronic intravascular haemolysis associated with prosthetic cardiac valves. Clin Sci 44: 511, 1973. 10. Leonardi P, Ruol P: Renal hemosiderosis in the hemolytic anemias: diagnosis by means of needle biopsy. Blood 16: 1029, 1960. 11. Hartmann RC, Butler SA, Jenkens DE, et al.: The kidneys in
hemoglobinuria should be evaluated for the presence of proximal tubular dysfunction; if renal tubular acidosis is present, therapy with appropriate doses of sodium bicarbonate should be instituted. ACKNOWLEDGMENT
We wish to thank the late Karl A. Stuart, M.D. who performed the renal biopsy.
PNH (abstract). Blood 44: 908, 1974. Crosby WH: Paroxysmal nocturnal hemoglobinuria. Relation to the clinical manifestations to undedying pathogenic mechanisms. Blood 8: 769, 1953. 13. Glaser RI: Haemoglobinurla and cardiovascular-renal disease. Am J Med 4: 594, 1948. 14. Rubin H: Paroxysmal hernoglobinuria with renal failure. JAMA 215: 433, 1971. 15. Strubing P: Paroxysmal haemoglobinuria. Dtsch Med Wochenschr 8: 17, 1882. 16. Jenkins DE: Diagnostic tests for paroxysmal nocturnal hemoglobinuria. Ser Haematol3: 24, 1972. 17. Wagley PF, Hickey MD: Susceptibility of red cells and serum factor in the mechanism of-hemolysis in paroxysmal nocturnal hemoglobinuria. J Clin Invest 27: 559, 1948. 18. Blum SF, Sullivan JM, Gardner FH: The exacerbation of hemolysis in paroxysmal nocturnal hemoglobinuria by strenuous exercise. Blood 30: 513, 1967. 19. Dameshek W, Fudenberg H: Paroxysmal nocturnal hemoglobinuric atypical manifestations suggesting an immunologic disease. Arch Intern Med 99: 202, 1957. 20. Ham T: Chronic hemolytic anemia with paroxysmal nocturnal hemoglobinuria. Study of the mechanism of hemolysis in relation to acid-base equilibrium. N Engl J Med 271: 915, 1937. Milne MD: Renal tubular dysfunction, chap 30. Diseases of 21. the Kidney (Strauss MB, Welt LG. eds), Boston, Little, Brown & co.. 1968. 12.
January 1977
The American Journal of Medicine
Volume 62
129
ARTHUR L. RILEY, M.D. LAWRENCE M. RYAN, M.D. DONALD A. ROTH, M.D. Milwaukee, Wisconsin
A patient with paroxysmal nocturnal hemoglobinuria, who required many blood transfusions for hemolytic episodes, had a perslstent hyperchloremic metabolic acidosis. Bicarbonate Infusion demonstrated a large fractional excretion of bicarbonate (28.8 per cent at a plasma bicarbonate level of 23 megAlter) whidr was consistent with proximal renal tubular acidosis. Generalized aminoaciduria and decreased tubular reabsorption of phosphate were also present. Marked deposition of iron in renal proximal tubules was associated with these functional abnormalities. We believe that, as systemic acidosis can promote hemolysis in patients with paroxysmal nocturnal hemogloblnuria, hemolysis can lead, by way of iron deposltion in renal tubules, to further acidosis. This cycle should be interrupted with appropriate doses of bicarbonate. Paroxysmal nocturnal hemoglobinuria is a rare disease characterized by the onset of hemolytic episodes in adults which are frequently worse at night and during intercurrent infections [ 11. Associated with this is a significant tendency to spontaneous venous thrombosis which may involve multiple organ systems but most frequently and dramatically involves abdominal viscera. We report a case of paroxysmal nocturnal hemoglobinuria complicated by renal tubular acidosis which most likely was caused by the paroxysmal nocturnal hemoglobinuria and, if untreated, potentially exacerbates the primary disorder. disease
CASE REPORT
From the Renal Section, Medical Service, Veterans Administration Center, Wood (Milwaukee), Wisconsin; and the Department of Medicine, The Medical College of Wisconsin, Milwaukee 53226. This study was supported in part by NIH Grant No. 5 MO 1 RR00058. Requests for reprints should be addressed to Dr. Arthur L. Riley, Renal Section/ 111, Veterans Administration Center, Wood, Wisconsin 53193. Manuscript accepted January 30, 1976.
This 31 year old white man first sought medical attention in 1966, with complaints of episodes of abdominal pain and vomiting for one year, associated with intermittent dark urine and fever. Laboratory data available from another hospital showed that the blood urea nitrogen was 12 mg/iOO ml, hematocrit value 26 per cent and reticulocyte count 19 per cent; 2+ proteinuria was present: the serum lactic dehydrogenase and indirect bilirubin levels were elevated. An x-ray series of the upper gastrointestinal tract revealed thickened jejunal folds which prompted an exploratory laparotomy. At surgery there were no discernible abnormalities of the liver, spleen or kidneys. A small, superficial jejunal ulcer was discovered and resected. After an uneventful, postoperative course the patient was discharged, with a diagnosis of fever of undetermined origin, jejunal ulcer and “hemolytic reaction.”
January 1977
The American Journal of Medicine
Volume 62
125
RENAL PROXIMAL TUBULAR DYSFUNCTION
I
TABLE
Time (min)
Effect of Bicarbonate Paroxysmal Nocturnal
Urine Volume (ml/mitt)
0
Urine pH 8.0
AND PAROXYSMAL NOCTURNAL HEMOGLOBINURIA-RILEY
Infusion on Per Cent Excretion of Filtered Bicarbonate Hemoglobinuria and Hyperchloremic Acidosis
Venous PH
7.27
5 per cent sodium bicarbonate 30 60 90 120
6.0 6.3 5.0 6.3
7.5 7.5 8.0 7.5
“Calculated from determination anaerobically.
ET AL.
7.35 7.37 7.38 7.37
Urine Bicarbonate* (meq/liter) 6.8
at 100 ml/hour 10.2 20.8 29.3 27.9
Plasma Bicarbonate* (meq/liter)
Urine Plasma Creatinine Creatinine ~mg/lOO ml) (mg/lOO ml)
19.1
(= 60 meq/hour) 20.4 23.2 22.4 23.4
of pH and of carbon dioxide pressure and total carbon dioxide
After discharge the patient continued to be anemic and to have episodes of abdominal pain. He was referred to the Mayo Clinic where the diagnosis of paroxysmal nocturnal hemoglobinuria was established by means of positive sugar-water and Ham tests. Subsequent treatment consisted of the administration of prednisone, 40 to 60 mg daily, and a transfusion of 2 to 4 lJ of buffy-coat-poor packed red cells (washed in saline solution) approximately every four months. The patient was first seen at the Veterans Administration Center, Wood, Wisconsin, in March 1971, because of ab-
in Patient
with
% Excretion of Filtered Bicarbonate
20
2.4
4.3
15 10 20 10
2.4 2.4 2.4 2.4
8.0 21.5 15.7 28.6
intravenously
by microgasometer
on specimens collected
dominal pain and hemolytic episodes. At that time, his blood urea nitrogen was 18 mg/lOO ml, sodium 140 meq/liter, potassium 5.7 meq/liter, chloride 108 meq/liter, carbon dioxide 20 meq/liter and hematocrit value 20 per cent. Urinalysis revealed a pH of 5.0, specific gravity of 1.014 and 1-F protein. At this time, striking hemosiderinuria was present in association with evidence of iron deficiency (low serum iron, high total iron-binding capacity, hypochromic microcytosis and absent marrow iron stores). By November 197 1, the serum creatinine was 1.2 mg/lOO ml and creatinine clearance 130 liters/day. A 24-hour urine specimen yielded
Figure 1. Left, photomicrograph of iron stain of kidney biopsy specimen, principally reveals marked iron deposition in tubules. Right, electron micrograph. Proximal tubular cell is identified by brush border; large amounts of electron-dense material are present within the cytoplasm. These deposits were not seen in cytoplasm of distal tubular cells. Magnification X 160 (left) and X 2,500 (right).
126
January 1977
The American Journal of Medicine
Volume 62
RENAL PROXIMAL
TUBULAR DYSFUNCTION
3 g of protein; serum bicarbonate was 13 meq/liter with a urine pH of 5.5. Oral bicarbonate, in doses of 6 to 8 g/day, was required to maintain a serum carbon dioxide level of 14 meq/liter or higher, and the urine pf-l was then consistently higher than 6.0. The patient received 82 U of buffy-coat-poor packed red blood cells (washed in saline solution) between March 1971 and February 1974, when he was readmitted for evaluation of renal dysfunction. No history of exposure to nephrotoxins could be elicited: Serum creatinine was now 2.3 mg/lOO ml and creatinine clearance 55 liters/day. Parathyroid hormone level was 74 pleq/ml (measured in the laboratory of Dr. Claude Arnaud, Mayo Clinic) with serum calcium 9.0 mg/lOO ml, serum phosphorous 4.4 mg/lOO ml and serum magnesium 2.14 mg/lOO ml. Serum protein electrophoresis was within normal limits. The plasma free-hemoglobin level was 423 mg/lOO ml and plasma haptoglobin level 120 mg/lOO ml. The urine specimen was strongly positive for occult blood. Although a 24-hour urine collection for the determination of protein (Biuret method) revealed 21 g, only 3 per cent of this was albumin. Electrophoresis demonstrated that the majority (61 per cent) was in the alpha2 zone with the rest in the beta (17 per cent) and gamma (17 per cent) zones; these three zones stained strongly positive on hemoglobin stain. Generalized aminoaciduria was identified; no significant glycosuria could be detected. Tubular reabsorption of phosphate was 50 per cent of the filtered load when the patient was adhering to a general diet. The patient was receiving 100 meq of sodium bicarbonate orally per day when a bicarbonate loading test was performed by infusion of 5 per cent sodium bicarbonate solution at 100 ml/hour (Table I). Before the infusion, the plasma carbon dioxide was 20 meq/liter, and 4.6 per cent of the filtered load of bicarbonate was excreted. When the serum carbon dioxide level reached 23.6 rneq/liter during the infusion, 28 per cent of the filtered load was excreted. Liver and skin biopsy specimens were obtained to evaluate hepatomegaly and hyperpigmentation. Hemosiderosis was identified in both tissues. An intravenous pyelogram showed radiopaque kidneys before injection of contrast media. Both kidneys were enlarged and measured 17 cm in length; the cortical outline was smooth and the calyceal system appeared normal. Renal biopsy tissue was grossly rust colored and, on microscopic examination, showed marked proximal tubular siderosis as well as mononuclear interstitial infiltrates and minimal proliferative changes in some glomeruli (Figure 1). lmmunofluorescence was negative for immunoglobulins G, A, M, complement and fibrin. Electron microscopy revealed marked deposition of electron-dense particles within the cytoplasm of proximal tubular cells (Figure 1). No such deposition was present in the distal tubular cells. Subsequently, the patient has had two episodes of abdominal pain during which radiopaque findings were consistent with venous infarction of the small bowel. On both admissions, with the patient taking supplemental bicarbonate, the urine pH ranged between 7.0 and 8.0. At present, the patient is maintained on a regimen of prednisone, 5 mg daily; folic acid, 1 mg daily: and 4 to 6 U of packed cells at intervals of two to four months. To maintain a plasma bicarbonate level
AND PAROXYSMAL
10
14
NOCTURNAL
18
SERUM
22
HEMOGLOBINURIA-RILEY
26
CARBON
30
3a
DIOXIDE
38
ET AL.
(meq/liter)
42
Fimre 2. Bicarbonate reabsorption and excretion in our patient (0) during the intravenous infusion of 3 per cent sodium bicarbonate. The normal response to bicarbonate infusion, taken from Pitts et al. [ 31, is shown for comparison. GFR = glomerular filtration rate. in the low normal range, therapy with 120 meq/day of sodium bicarbonate is required. At no time has this patient demonstrated a nocturnal pattern of hemoglobinuria. COMMENTS
The patient had ample evidence of renal tubular acidosis, as demonstrated by a persistently alkaline urine in the presence of a metabolic hyperchloremic acidosis, but no evidence of a urinary tract infection. The identification of the acidosis as a proximal tubular defect was made by the abnormally low renal threshold for bicarbonate, with 4 per cent excretion of the filtered load at a serum concentration of 20 meq/liter and more than 20 per cent excretion at serum levels of 24 meq/liter [2]. The patient also required large amounts of bicarbonate supplement to increase the serum carbon dioxide to 20 meq/liter. In Figure 2, the responses to bicarbonate infusion in our patient are compared with the normal response to bicarbonate infusion [3]. The results suggest not only an early threshold for bicarbonate excretion but also a reduced capacity for bicarbonate reabsorption. Other proximal defects identified included aminoaciduria and decreased tubular reabsorption of phosphate (50 per cent during the adherence to a general diet). Recently, Muldowney et al. [4] demonstrated a reduction in maximal tubular reabsorptive capacity for bicarbonate in 15 patients with chronic renal failure of various etiologies with creatinine clearance ranging from 2.7 to 12.0 ml/min; most of these patients suffered either from pyelonephritis or from polycystic disease and had been uremic for relatively long periods. The
January 1977
The American Journal of Medicine
Volume 62
127
RENAL
PROXIMAL
TUBULAR
DYSFUNCTION
AND
PAROXYSMAL
NOCTURNAL
defects seemed related to the duration and degree of secondary hyperparathyroidism. Aminoaciduria was absent. It is unlikely that the proximal tubular defects in our patient can be explained only by “parathyroid acidosis” because (1) serum bicarbonate values were decreased and serum chloride values were increased in 1971 before the development of azotemia (creatinine clearance was 130 liters/day); (2) in 1974, the parathyroid hormone level was predictably high, but the mild to moderate azotemia at the time of bicarbonate infusion had been present less than one year; (3) the presence of aminoaciduria in our patient favors a more generalized proximal tubular disorder; and (4) the presence of marked iron deposition in the proximal tubules suggests that the defect in proximal tubular function is related to this morphologic lesion. Heavy metals are known to result in kidney damage, with manifestations which can include acute oliguric renal failure, chronic renal failure, the nephrotic syndrome and, importantly, abnormalities in proximal tubular function manifested by renal aminoaciduria, renal glycosuria and excessive loss of phosphate. Hyperchloremic acidosis, without further definition, is also described. Pathologic changes in epithelial cells of the proximal tubule correlate with this functional impairment
ia. A prominent part of our patient’s disease has been hemosiderosis as confirmed by liver, skin and renal biopsy. Whereas deposition of iron in the proximal tubules (and in the urine sediment) of patients with paroxysmal nocturnal hemoglobinuria is present even in the face of severe iron deficiency [6], the significant hemosiderosis in other tissues in this patient is the result of multiple transfusions. The deposition of iron pigment in the kidney was preferential for the proximal tubules, providing an anatomic correlate to his physiologic defect. Iron deposition in proximal tubules in various states HEMOLYSI / SYSTEMIC AFIDOSIS
BICARBONATE LOSS IN URINE
S t IRON DEPOSITION RENAL TUBULES (PR~XIMAL~
IN I
PROXIMAL
I
WRTA Vgure 3. Proposed scheme of interrelationship between paroxysmal nocturnal hemoglobinuria and renal tubular acidosis. The administration of appropriately large doses of bicarbonate can interrupt this cycle.
120
January 1977
The American Journal of Medicine
HEMOGLOBINURIA-RILEY
ET AL.
of intravascular hemolysis has been well described in the past [7-lo], but evidence of a concomitant functional defect has been lacking. In only one recent preliminary communication were specific defects of tubular function in patients with paroxysmal nocturnal hemoglobinurla considered; hyposthenuria was found in nine and abnormal phenolsulfonphthalein clearance in seven of the nine patients tested [ 111. Azotemia has occasionally been seen in patients with paroxysmal nocturnal hemoglobinuria, and chronic pyelonephritis [ 121, nephrosclerosis [ 131 or transfusion reactions [ 141 have been implicated. The significance of renal tubular acidosis in patients with paroxysmal nocturnal hemoglobinuria lies in the effect of systemic acidosis upon hemolysis. Strubing [ 151, in the first clinical description of paroxysmal nocturnal hemoglobinuria, noted the striking relationship of hemoglobinuria to sleep, attributing this to an accumulation of carbon dioxide or lactic acid. The Ham test (acid hemolysin test) is an in vitro determination, nearly specific for paroxysmal nocturnal hemoglobinuria, in which a pH of 6.5 augments hemolysis [ 161. Less dramatic alterations of pH, from 7.4 to 7.2 in vitro, also result in increased hemolysis [ 171. Local acidosis induced by arm exercise has been show to increase local hemolysis in a patient with paroxysmal nocturnal hemoglobinuria, associated with a decrease in antecubital vein pH from 7.48 to 7.33 [ 181. The administration of acetazolamide has had a similar effect [ 191. Finally, a large single dose of ammonium chloride has been shown to increase hemoglobinemia and hemoglobinuria with a simultaneous decrease of the arterial pH to 7.25 [20]. Therefore, it is reasonable to conclude that the lowering of blood pH in vivo can exacerbate paroxysmal nocturnal hemoglobinuria. One hypothesized mechanism is that acidosis enhances the sensitivity of paroxysmal nocturnal hemoglobinuria cells to complement. Our patient presented with paroxysmal nocturnal hemoglobinuria and coexistent renal tubular acidosis, aminoaciduria, proteinuria and decreased glomerular filtration. The biopsy finding of iron deposition, largely in the proximal convoluted tubules, correlates well with the functional tubular defects. In our review of the literature on paroxysmal nocturnal hemoglobinuria, no description of renal tubular acidosis was found. However, since the urine of patients with proximal renal tubular disease can be acidified when serum bicarbonate levels are very low, it is possible that the renal origin of such acidosis could be overlooked unless specifically pursued. The etiology of the tubular defect may well be related to the particular pattern of iron deposition seen, just as deposition of other metals, such as lead, uranium, cadmium and copper [21], may cause a spectrum of proximal tubular dysfunctions.
Volume 62
RENAL PROXIMAL TUBULAR DYSFUNCTION AND PAROXYSMAL NOCTURNAL HEMOGLOBINURIA-RILEY ET AL.
The therapeutic implications are important in that, as paroxysmal nocturnal hemoglobinuria may induce tubular changes, the resulting acidosis can accelerate hemolysis and lead to a self-perpetuating cycle of hemolysis causing proximal tubular dysfunction which increases acidosis and further hemolysis (Figure 3). Such a pathway can be interrupted by administering bicarbonate in large enough quantities to correct the metabolic acidosis. Patients with paroxysmal nocturnal
Dacie JV, Lewis SM: Paroxysmal nocturnal haemoglobinuria: clinical manifestations, haematology, and nature of the disease. Ser Haematol 3: 3, 1972. 2. Morris RC: Renal tubular acidosis: mechanisms, classification and implications. N Engl J Med 281: 1405, 1969. 3. Pitts RF, Ayer JL, Schiess WA: The renal regulation of acidbase balance in man. III. The reabsorption and excretion of bicarbonate. J Clin Invest 28: 35, 1949. 4. Muldowney FP, Donohoe JF, Carroll DV, et al.: Parathyroid acidosis in uraemia. Q J Med 41: 321. 1972. 5. Emerson BT: Metals and the kidney, chap 23. Renal Disease (Black DAK, ed), Philadelphia, F. A. Davis Co., 1967. 6. Hutt AT, Roger JF, Neustein HB: Renal pathology in paroxysmal nocturnal hemoglobinuria. Am J Med 31: 736, 1961. 7. Lamarca JT, Terroba JY, Diaz-Flores L: Hemoglobinuria paroxistica noctuma. II. lnsuficiencia Renal en el Curso de la Misma. Rev Clin Esp 28: 287, 1973. 8. Sigler AT, Furman EN, Zinkham WH, et al.: Severe intravascular hemolysis following surgical repair of endocardial cushion defects. Am J Med 35: 467, 1963. 9. Slater SD, Rahman M, Lindsay RM: Renal function in chronic intravascular haemolysis associated with prosthetic cardiac valves. Clin Sci 44: 511, 1973. 10. Leonardi P, Ruol P: Renal hemosiderosis in the hemolytic anemias: diagnosis by means of needle biopsy. Blood 16: 1029, 1960. 11. Hartmann RC, Butler SA, Jenkens DE, et al.: The kidneys in
hemoglobinuria should be evaluated for the presence of proximal tubular dysfunction; if renal tubular acidosis is present, therapy with appropriate doses of sodium bicarbonate should be instituted. ACKNOWLEDGMENT
We wish to thank the late Karl A. Stuart, M.D. who performed the renal biopsy.
PNH (abstract). Blood 44: 908, 1974. Crosby WH: Paroxysmal nocturnal hemoglobinuria. Relation to the clinical manifestations to undedying pathogenic mechanisms. Blood 8: 769, 1953. 13. Glaser RI: Haemoglobinurla and cardiovascular-renal disease. Am J Med 4: 594, 1948. 14. Rubin H: Paroxysmal hernoglobinuria with renal failure. JAMA 215: 433, 1971. 15. Strubing P: Paroxysmal haemoglobinuria. Dtsch Med Wochenschr 8: 17, 1882. 16. Jenkins DE: Diagnostic tests for paroxysmal nocturnal hemoglobinuria. Ser Haematol3: 24, 1972. 17. Wagley PF, Hickey MD: Susceptibility of red cells and serum factor in the mechanism of-hemolysis in paroxysmal nocturnal hemoglobinuria. J Clin Invest 27: 559, 1948. 18. Blum SF, Sullivan JM, Gardner FH: The exacerbation of hemolysis in paroxysmal nocturnal hemoglobinuria by strenuous exercise. Blood 30: 513, 1967. 19. Dameshek W, Fudenberg H: Paroxysmal nocturnal hemoglobinuric atypical manifestations suggesting an immunologic disease. Arch Intern Med 99: 202, 1957. 20. Ham T: Chronic hemolytic anemia with paroxysmal nocturnal hemoglobinuria. Study of the mechanism of hemolysis in relation to acid-base equilibrium. N Engl J Med 271: 915, 1937. Milne MD: Renal tubular dysfunction, chap 30. Diseases of 21. the Kidney (Strauss MB, Welt LG. eds), Boston, Little, Brown & co.. 1968. 12.
January 1977
The American Journal of Medicine
Volume 62
129
![[Paroxysmal nocturnal hemoglobinuria].](/assets/img/hold.png)
