Prospective randomized trial of maintenance immunosuppression with rapid discontinuation of prednisone in adult kidney transplantation.

Rapid discontinuation of prednisone (RDP) has minimized steroid-related complications following kidney transplant (KT). This trial compares long-term (10-year) outcomes with three different maintenance immunosuppressive protocols following RDP in adult KT. Recipients (n=440; 73% living donor) from March 2001 to April 2006 were randomized into one of three arms: cyclosporine (CSA) and mycophenolate mofetil (MMF) (CSA/MMF, n=151); high-level tacrolimus (TAC, 8-12 µg/L) and low-level sirolimus (SIR, 3-7 µg/L) (TACH/SIRL, n=149) or low-level TAC (3-7 µg/L) and high-level SIR (8-12 µg/L) (TACL/SIR(H) , n=140). Median follow-up was ∼7 years. There were no differences between arms in 10-year actuarial patient, graft and death-censored graft survival or in allograft function. There were no differences in the 10-year actuarial rates of biopsy-proven acute rejection (30%, 26% and 20% in CSA/MMF, TACH/SIRL and TACL/SIRH) and chronic rejection (38%, 35% and 31% in CSA/MMF, TACH/SIRL and TACL/SIRH). Rates of new-onset diabetes mellitus were higher with TACH/SIRL (p=0.04), and rates of anemia were higher with TACH/SIRL and TACL/SIRH (p=0.04). No differences were found in the overall rates of 16 other post-KT complications. These data indicate that RDP-based protocol yield acceptable 10-year outcomes, but side effects differ based on the maintenance regimen used and should be considered when optimizing immunosuppression following RDP.

The glomerular mesangium. 3. Acute immune mesangial injury: a new model of glomerulonephritis.

A mechanism of immune glomerular injury is described based on the fixation of antibody (Ab) to an antigen (Ag) that has localized in the glomerular mesangium. Rabbits were given, intravenously (i.v.), aggregated human IgG (AHIgG) or albumin (AHSA) and 10 h later, when the Ag by immunofluorescent microscopy was present in the mesangium, a kidney was removed and transplanted into a normal rabbit. The recipient then received, i.v., rabbit anti-HIgG or anti-HSA. Within minutes of Ab infusion, glomeruli of the donor kidney had polymorphonuclear (PMN) infiltration that over the next few hours became marked and was associated with glomerular cell swelling. At 24 h a decrease in PMN's and early mesangial proliferation was seen. By 3 days there was marked mesangial hypercellularity and increased mesangial matrix. Within minutes after Ab administration rabbit IgG, C3, and fibrin were seen in the glomerular mesangium. There was a fall in complement titer by 1 min after Ab infusion that was due to complement consumption by the donor kidney. Complement then returned to normal levels by 48 h. Significant glomerular injury did not occur (a) in the recipient's own kidney, (b) from Ag administration and transplantation without recipient Ab administration, or (c) from transplantation and Ab administration without prior Ag administration. These studies demonstrated that Ag localized in the glomerular mesangium can react with circulating Ab and complement resulting in severe glomerular injury.

Renal transplantation in diabetes mellitus in rats.

Immunoglobulins and complement are deposited in the glomerular mesangium of rats with progressive glomerulosclerosis secondary to chemically induced diabetes mellitus. Isotransplantation of a kidney from a rat diabetic for 6 mo into a normal recipient results within 2 mo in the disappearance of IgG, IgM, and beta(1)C from the mesangium and arrest or reversal of the light microscopic glomerular lesions. Kidneys isotransplanted from normal donors into diabetic rats developed mesangial matrix thickening and deposition of IgG, IgM) and beta(1)C in the mesangium. No glomerular changes occur upon transplantation of a normal kidney into a normal rat. These findings indicate that diabetic glomerular changes in the rat are reversible and are secondary to the diabetic state rather than to the inducing agent.

H-2 antigen class: effect on mouse islet allograft rejection.

Rejection of mouse pancreatic islet allografts occurred in a high percentage of donor recipient combinations identical for H-21-region antigens and differing at H-2K and H-2K + H-2D without I-region disparities. The results suggest that disparities in major histocompatibility complex antigens of class I (H-2K and H-2D) alone are capable of eliciting islet allograft rejection, and that lack of a stimulus from class II (I-region) alloantigens does not ensure permanent islet allograft survival.

Hepatocellular transplantation for metabolic deficiencies: decrease of plasms bilirubin in Gunn rats.

A sustained decrease of plasma bilirubin concentrations occurred in homozygous recessive Gunn rats lacking the enzyme uridine diphosphate glucuronyltransferase following infusion into the portal vein of hepatocytes from heterozygous nonjaundiced Gunn rats possessing the enzyme. Transplantation of cells capable of continuous enzyme production could be an effective mode of therapy for congenital enzyme deficiency diseases.

2202 kidney transplant recipients with 10 years of graft function: what happens next?

The ultimate goal of clinical transplantation is for the recipients to achieve long-term survival, with continuing graft function, that is equivalent to that of the age-matched general population. We studied subsequent outcome in kidney transplant recipients with 10 years of graft function. In all, 2202 kidney transplant recipients survived with graft function >10 years. For 10-year survivors, the actuarial 25-year patient survival rate for primary transplant living donor (LD) recipients was 57%; graft survival, 43%. For primary transplant deceased donor (DD) recipients, the actuarial 25-year patient survival rate was 39%; graft survival, 27%. The two major causes of late graft loss were death (with graft function) and chronic allograft nephropathy (tubular atrophy and interstitial fibrosis). The two major causes of death with function were cardiovascular disease (CVD) and malignancy. For nondiabetic recipients, the mean age at death with function from CVD was 54 +/- 13 years; for diabetic recipients, 53 +/- 7 years. By 20 years posttransplant, morbidity was common: >40% recipients had skin cancer (mean age for nondiabetic recipients, 53 +/- 13 years; for diabetics, 49 +/- 8 years), >10% had non-skin cancer (mean age for nondiabetic recipients, 53 +/- 16 years; for diabetics, 46 +/- 9 years), and >30% had CVD (mean age for nondiabetic recipients, 53 +/- 15 years; for diabetics, 47 +/- 9 years). We conclude that long-term transplant recipients have a high rate of morbidity and early mortality. As short-term results have improved, more focus is needed on long-term outcome.

Clinical spectrum of lymphoproliferative disorders in renal transplant recipients and evidence for the role of Epstein-Barr virus.

Six renal transplant recipients with abnormal lymphoproliferative disorders were studied in an attempt to define their clinical features and the role of Epstein-Barr virus (EBV) in their pathogenesis. Patients were either teenage (three) or in the sixth decade (three). The younger patients presented an average of 3 months after transplantation with fever, sore throat, and lymphadenopathy; had been markedly immunosuppressed; frequently had preceding or concomitant cytomegalovirus infections; and two of three had a rapidly fatal course. The older patients presented an average of 5 years after transplantation while on maintenance immunosuppressive drugs; in two of three cases with an oropharyngeal tumor; and had a more indolent, but frequently fatal, clinical course. The most frequent sites of biopsy-proven involvement in these patients were lymph nodes (three), the oropharynx (three), liver (three), bone marrow (three), transplanted kidney (three), colon (two), and central nervous system (two). EBV-specific antibody titers including anti-viral capsid antigen IgG, anti-viral capsid antigen IgM, anti-early antigen, and anti-Epstein-Barr nuclear antigen were serially measured in all patients. Four patients demonstrated serological evidence of a primary (one) or reactivation (three) EBV infection. No patient had significant changes in anti-early antigen or anti-Epstein-Barr nuclear antigen titers. All three patients tested for oropharyngeal shedding of EBV were positive. A touch imprint of one tumor was stained for the presence of Epstein-Barr nuclear antigen, and a majority of cells were positive. EBV complementary RNA/DNA filter hybridization and/or viral DNA/DNA reassociation analysis performed on tumor biopsy specimens in five patients demonstrated multiple EBV genome equivalents per cell in all eight specimens tested. Clinical, pathological, serological, and molecular hybridization studies provide substantial evidence that EBV was the cause of these lymphoproliferative disorders occurring after renal transplantation. Impaired host defenses allow the EBV-transformed B-lymphocytes to escape normal control mechanisms. This impairment is invariable and influenced by many factors resulting in the observed spectrum of disease. Cytogenetic changes, however, may also be important.

Metabolic effect of hemipancreatectomy in donors. Preoperative prediction of postoperative oral glucose tolerance.

Between November 1977 and January 1987, 55 transplantations of pancreas segments from living donors related to their recipients were performed at the University of Minnesota. A preliminary analysis of metabolic test results in donors tested 1 yr after hemipancreatectomy showed an increase in mean glucose and a decrease in mean insulin values during oral glucose tolerance tests (OGTTs) in 18 donors, a 14% increase in the mean of the mean glucose levels during 24-h metabolic profiles in 12 donors, and a decrease of 45% in the mean 24-h urinary C-peptide excretion in 21 donors. Including the studies performed postdonation, 11 of 31 (35%) donors developed an abnormal OGTT result. In a retrospective analysis, preoperative results of intravenous glucose tolerance tests (IVGTTs) and cortisone-stimulated OGTTs were found to be statistically significant predictors of an abnormal OGTT after hemipancreatectomy. The mean of the 5- to 50-min IVGTT insulin values was the best predictive test. With the cutoff value set at 62 microU/ml, this test result had a sensitivity of 100%, a specificity of 83%, and a positive predictive value of 75% for identifying donors who developed an abnormal OGTT. The sum of the 5- and 10-min IVGTT insulin (cutoff 140 microU/ml) had a sensitivity of 100%, a specificity of 67%, and a predictive value of only 60%, whereas the delta-insulin had values of 86, 71, and 60%, respectively. Both the IVGTT mean insulin and the sum of the 5-min and 10-min insulin test results were 100% predictive of an abnormal test (0% risk), but the IVGTT mean insulin excluded the lowest proportion of otherwise suitable donors (a low "false-alarm" rate). The IVGTT mean insulin can be used to identify or exclude potential donors who would develop an abnormal OGTT result should hemipancreatectomy be performed.

Clinical segmental pancreas transplantation without duct anastomosis in diabetic renal allograft recipients.

Segmental pancreatic allografts with unligated ducts were transplanted intraperitoneally to five diabetic patients who had received renal allografts more than 2 yr earlier. One patient is alive with a functioning graft 10.5 mo later. Two patients rejected their grafts at approximately 2 and 3 mo but are alive 8--9 mo later after resumption of exogenous insulin therapy. In both patients, carbohydrate metabolism was nearly normal during the period of graft function. Two patients died of infectious complications between 1 and 2 mo after transplantation. The main hazard of pancreas transplantation relates to the immunosuppression necessary to prevent rejection, setting the stage for infectious complications. Technically, pancreas transplantation is a feasible and efficient procedure, and, if better methods are developed for preventing rejection, it should be applicable to patients prone to develop secondary complications of diabetes.

Transplantation of dispersed pancreatic islet tissue in humans: autografts and allografts.

Islet transplantation is successful in animals and holds considerable promise as endocrine replacement therapy for patients with diabetes mellitus, but clinical application to diabetic patients has been difficult. We have shown the technical feasibility of human islet transplantation by autotransplantation of dispersed pancreatic islet tissue into the portal vein in three patients with chronic pancreatitis and incapacitating, intractable pain who underwent near-total (greater than 97%) pancreatectomy. In all three patients, the excised pancreas was dispersed by collagenase digestion, but no effort was made to purify the islets. Islet yield, as judged by tissue insulin content, ranged from 24 to 55%. The first patient, who never received insulin after the pancreatectomy and islet autotransplantation, had a normal oral glucose tolerance test by 3 wk and has remained normoglycemic for over 2 yr. In the second patient, viable islets were histologically identified in the liver parenchyma. The third patient was treated with hyperalimentation for 3 wk after the pancreatectomy and islet autotransplantation and, during this period, required insulin. After cessation of hyperalimentation and initiation of oral geedings, the patient was withdrawn from insulin. Although abnormalities of carbohydrate metabolism were present, the patient did not require insulin for more than 1 yr. Seven diabetic renal allograft recipients have received allografts of dispersed pancreatic islet tissue prepared in the same way. No patients were cured of diabetes, although transient evidence of islet function--increase in serum or urinary C-peptide levels or decrease in exogenous insulin requirements--occurred in some. Although rejection was probably responsible for most of the failures, transplantation of allogeneic human islet tissue as a free graft is metabolically inefficient. With the current state of immunosuppressive therapy, the primary role of islet transplantation may be in a situation where rejection cannot occur: as an autograft to obviate the occurrence of diabetes after extensive pancreatectomy for benign disease.

Infant human pancreas. A potential source of islet tissue for transplantation.

Twelve pancreases from human infants one year old or less were analyzed for tissue insulin and amylase content before and after dispersal of pancreatic fragments by mincing and collagenase digestion. Tissue insulin and amylase content provide an index of pancreatic islet mass and exocrine digestive enzyme content, respectively. The results were compared with similar anaylses performed on juvenile and adult human pancreases before and after islet isolation and on intact and dispersed neonatal rat and adult rat pancreas. Infant human pancreas has an average tissue insulin concentration of 1,128 mug./gm. of tissue and a total insulin content of 1,718 mug/pancreas, as against values of 140 mug./gm. of tissue and 7,209 mug./pancreas for adult human pancreas. Average tissue amylase concentration is 0.24 mg./gm. of tissue in infant human pancreas and 3.0 mg./gm. of tissue in adult human pancreas. The insulin/amylase ratio in infant pancreas is 4,800, as against 46 in the adult pancreas. Neonatal rat pancreas, which can be dissociated and transplanted without separation of islet and exocrine components, has a similarly high tissue insulin and low tissue amylase content when compared with adult rat pancreases. Infant human pancreas has a total islet mass 24 per cent that of an adult human pancreas, and neonatal rat pancreas has a total islet mass 11 per cent of that of an adult rat pancreas. One neonatal rat pancreas prepared by minimal collagenase digestion can cure diabetes when transplanted via the portal vein to a rat. Following dispersal of infant human pancreas by collagenase digestion, the islet content and the insulin/amylase ratio of the recovered tissue equals or exceeds that which usually can be isolated from adult cadaver pancreases. Infant human pancreas is a rich source of islet tissue that is relatively uncontaminated by exocrine digestive enzymes. After dispersal, infant human pancreas may be ideal for transplantation to selected diabetic patients.

Comparison of two methods of islet preparation and transplantation in dogs.

Seventy-nine mongrel dogs underwent total pancreatectomy. Fifteen dogs served as apancreatic controls and died 7.0 +/- 4.2 days later (mean +/- SD). The pancreases of 44 dogs (group 1) were intraductally distended by manual injection of Hanks' balanced salt solution (HBSS). Thereafter each organ was mechanically disrupted and subjected to collagenase digestion as described by Mirkovitch et al. The pancreases of 20 dogs (group 2) were intraductally distended and subsequently perfused with collagenase by a roller pump. The organs were then mechanically disrupted and filtered through screens as described by Horaguchi et al. The resulting tissue suspensions were injected into the spleens of the dogs as autotransplants in both groups, by direct punction of the splenic capsule in group 1 and by retrograde infusion via a splenic vein tributary in group 2. The functional outcome was better in group 2 than in group 1, as assessed by the number of animals that became normoglycemic after transplantation [15/20 (75%) vs. 13/44 (30%); P = .0025]. The degree of islet purification, as measured by an increase in the tissue insulin/amylase ratio, was higher in group 2, and in both groups it was higher in normoglycemic than in hyperglycemic animals. The percent engraftment [i.e., amount of insulin recovered from spleen as percent of tissue transplanted (mean, 15.4% in group 1 and 14.5% in group 2) or as percent of original pancreas (mean, 4.9% in group 1 and 4.4% in group 2)] was low in both groups but again was higher in normoglycemic than in hyperglycemic animals within each group. In conclusion, both the degree of engraftment and purification and the route of implantation influenced the functional outcome after dispersed pancreatic islet autotransplantation to the spleen of totally pancreatectomized dogs, with purified tissue injected retrogradely functioning better than unpurified tissue injected directly.

Pancreas-transplant outcome in relation to presence or absence of end-stage renal disease, timing of transplant, surgical technique, and donor source.

The differences in pancreas-transplant outcome according to recipient status, surgical approach, and donor source are illustrated by an analysis of results at one institution with experience in several categories. From July 1978 to January 1988, 210 pancreas transplants were performed, and 67 grafts are still functioning, the longest for 9.7 yr. Since October 1984, a uniform immunosuppressive protocol has been used, antilymphocyte globulin, cyclosporin, azathioprine, and prednisone for induction and the last three drugs for maintaining antirejection therapy. During this period, 110 pancreas transplants were performed, 62 in nonuremic non-kidney transplants, 28 in recipients of a previous kidney, and 20 simultaneous with a kidney; 64 with bladder and 43 with enteric drainage; and 25 from related and 85 from cadaver donors. The overall patient survival rate at 1 yr was 91%, and there were no significant differences between the various categories. Graft survival rates, however, differed between the various categories created by combinations of the above variables. With bladder drainage, 1-yr function rates were 58% (n = 30), 47% (n = 15), and 77% (n = 19) in recipients of a pancreas transplant alone, a pancreas after a kidney, or a simultaneous pancreas-kidney transplant; with enteric drainage, 1-yr function rates were 33% (n = 32) and 36% (n = 11) in the pancreas transplant alone and pancreas after kidney categories (enteric drainage was not done in double-transplant patients).(ABSTRACT TRUNCATED AT 250 WORDS)

The development of lesions in the glomerular basement membrane and mesangium after transplantation of normal kidneys to diabetic patients.

Renal allograft biopsies at the time of transplantation (baseline) and 2 yr later were obtained in 6 type I diabetic and 12 nondiabetic patients and studied for glomerular basement membrane (GBM) and mesangial changes. Diabetic patients had significantly greater GBM thickness compared with nondiabetics at 2 yr (P = 0.05, rank sum test), and the increase in GBM thickness comparing baseline and 2-yr biopsies was greater in the diabetic compared with nondiabetic patients (P = 0.005, rank sum test). Similarly, diabetic patients developed significant mesangial thickening by light microscopy while no changes were observed in nondiabetic patients (P = 0.001). Electron microscopic morphometric analysis of the percentage of total mesangium was not different on comparing diabetic and nondiabetic patients at 2 yr. There was an increase in the matrix component of the mesangium in the diabetics at this time, although this did not reach statistical significance (P = 0.06). In addition, the surface density of the peripheral glomerular capillary wall, presumably reflecting mesangial expansion, was decreased in the diabetic and unchanged in the nondiabetic patients (P = 0.005). These studies document, for the first time, the development of GBM and mesangial lesions of diabetic nephropathy in normal living related donor and cadaver kidneys transplanted into diabetic patients and support the hypothesis that these lesions are secondary to the diabetic state.

Pregnancy in a juvenile diabetic after renal transplantation (class T diabetes mellitus).

The successful outcome of a pregnancy in a juvenile diabetic after renal transplantation is reported. It is proposed that class T be added to the classification of pregnancies complicated by diabetes mellitus. Pregnancy prevention should be considered until significant longevity can be demonstrated in diabetics receiving renal transplants.

Long-term follow-up of polyneuropathy in diabetic kidney transplant recipients.

Nerve conduction and electromyography (EMG) of insulin-dependent (type 1) diabetic patients with end-stage nephropathy was studied before and up to 10 yr after kidney transplantation (KTx). A series of nondiabetic KTx patients served as a comparison group. Motor nerve conduction velocity (NCV) was measured in the ulnar, median, peroneal, and tibial nerves; sensory NCV was measured in the median nerve. EMG was performed in the first dorsal interosseus, flexor carpi radialis, anterior tibialis, and gastrocnemius muscles. In 68 pre-KTx diabetic patients, the mean NCV was below normal in all nerves, and the mean amplitudes of the evoked muscle action potential (MAP) were low normal in the upper extremity and below normal in the lower extremity. The values of the comparison group were within the normal range. At 1 (n = 57), 5 (n = 23), and 10 (n = 10) yr after KTx, the mean NCV of the diabetic patients remained essentially unchanged, but MAP amplitudes of all muscles had declined. EMG revealed progression of the denervation process, especially in muscles of the lower extremities. We conclude that diabetic neuropathy continues to progress by a progressive axonal loss after correction of uremia by KTx.

Immunopathology of renal extracellular membranes in kidneys transplanted into patients with diabetes mellitus.

Kidneys of patients with severe diabetic nephropathy demonstrate marked linear immunofluorescent staining of extracellular membranes, including the tubular and glomerular basement membranes (TBM and GBM) and Bowman's capsule. Immunofluorescent studies were carried out on kidney tissue obtained from 12 diabetic and 17 nondiabetic patients from two to 12 years following renal transplantation. The frequency and intensity of SgG and albumin staining of these membranes were significantly greater in the diabetic than in the nondiabetic patients (P less than 0.0005). TBM, GBM, and Bowman's capsule staining did not occur in any of the seven kidneys studies at the time of their transplantation into diabetic recipients. Thus, the abnormalities leading to the deposition or trapping of proteins in renal extracellular membranes occur early after the placement of normal kidneys into the abnormal metabolic environment of the diabetic transplant recipient. The present study supports the concept that basement membrane alterations in diabetes are a consequence of the biochemical perturbations of diabetes rather than a separately inherited genetically linked disorder.

Cyclosporin A for immunosuppression: observations in rat heart, pancreas, and islet allograft models and in human renal and pancreas transplantation.

Cyclosporin A (CsA) is a unique immunosuppressive cyclic polypeptide that is currently being used, either alone or in combination with low-dose prednisone, to treat recipients of renal or pancreas allografts in clinical trials. CsA is very effective in preventing rejection of heart and renal allografts in rodents, but in nontoxic doses does not consistently prevent rejection of pancreas and islet allografts. Therefore, we tested low-dose CsA in various combinations with low-dose prednisone, azathioprine, or total lymphoid irradiation in rat heart, pancreas, and islet allograft models. Several combinations are synergistic and when administered continuously can indefinitely prevent rejection of heart allografts, but only delay rejection of pancreatic allografts, transplanted across a major histocompatibility barrier, CsA by itself prolonged the survival of islet allografts transplanted across a minor, but not a major, histocompatibility barrier. CsA and azathioprine had a synergistic effect in the minor histocompatibility barrier islet transplant model, but, in the nontoxic combinations tested, could not prevent rejection indefinitely. A randomized prospective trial comparing standard immunosuppressive therapy (ALG, prednisone, and azathioprine), with CsA and low-dose prednisone for clinical renal allotransplantation is ongoing at the University of Minnesota. Current actuarial 1-yr graft survival is 93% for CsA-treated patients (N = 48) and 81% for conventionally treated patients (N = 52). Patient survival is 98% for CsA and 100% for conventionally treated patients. A pilot trial of CsA in the clinical pancreas transplant program at the University of Minnesota is also underway. Since 1978, 46 pancreas transplants have been performed in 43 patients. Of 30 technically successful pancreatic allografts, 5 of 12 recipients treated with conventional immunosuppression and 6 of 18 recipients treated with CsA currently have functioning grafts and are insulin independent between 1 and 44 months after transplantation. The results of metabolic studies are similar in conventional and CsA-treated patients with functioning pancreas grafts. Since pancreas grafts may fail for reasons other than rejection, further observations are needed to ascertain the role of CsA in clinical pancreas transplantation.

Long-term study of normal kidneys transplanted into patients with type I diabetes.

We examined the rate of development of the lesions of diabetic nephropathy in transplanted kidneys residing for 6-14 yr in type I (insulin-dependent) diabetic kidney-allograft recipients. Although each recipient had end-stage diabetic nephropathy with his/her own kidneys, there was marked variability in the rate of development of mesangial expansion observed in the transplanted kidneys. The progression of glomerular pathology, including widening of the glomerular basement membrane and expansion of the mesangium, did not correlate significantly with several potential risk factors (e.g., donor source--cadaver or living related, histocompatibility match, age of the recipient or donor, age at onset of diabetes, duration of diabetes before renal failure, immunosuppressive drug dose, blood pressure, and severity of lesions of chronic rejection). However, there was a direct albeit imprecise relationship between the index of mesangial expansion at the final biopsy and the index of glycemic control (r = .61, P less than .01). These observations suggest that currently unknown factor(s) may modulate the progression of diabetic renal disease, even in a population that had uniformly demonstrated nephropathy risk. Our data support the hypothesis that in addition to glycemic control per se, there are risk factors intrinsic to the kidney itself.

Mycobacterial infections in renal transplant recipients. Seven cases and a review of the literature.

During an 11-year period, 1,069 patients received renal allografts at the University of Minnesota Hospital, Minneapolis, and infections developed in seven (0.65%) due to mycobacteria (Mycobacterium tuberculosis and M kansasii). The primary infection was in joint or subcutaneous tissue in six patients and pulmonary (miliary) in one. Infections in joint or skin shared common features regardless of the species of Mycobacterium and usually mimicked acute pyogenic bacterial infection; all responded to antimycobacterial drugs. The clinical manifestations in our patient in miliary tuberculosis were compared with those of 19 other patients described in the literature. Although their systemic manifestations were more severe, the symptoms were often ill-defined and the diagnosis overlooked. Five of these 20 patients (25%) died of uncontrolled infection.