Anti-glomerular basement membrane (GBM) disease (Goodpasture’s disease) is a rare autoimmune disorder characterised by rapidly progressive glomerulonephritis (RPGN) with crescentic changes affecting most glomeruli on renal biopsy, and often accompanied by pulmonary haemorrhage. The disease is, by definition, associated with the development of anti-GBM autoantibodies. The incidence is estimated to be 0.5 to 1 new case per million population per year, and it accounts for up to 7% of patients with end stage renal failure, and 1 to 2% of all renal biopsies. Most patients are Caucasoid. There is a bimodal age distribution with peaks occurring in the third and sixth decades, and a slight excess of men.

Clinical presentation

Most patients present acutely with lung haemorrhage, RPGN, or both. Lung haemorrhage tends to lead to early presentation, as it is a particularly alarming symptom and occurs in about two thirds of patients, while renal failure even in the context of RPGN may be insidious. However, the spectrum of disease is wide; some patients have chronic recurrent alveolar haemorrhage manifest only as iron deficiency, a few patients report recurrent minor episodes of haemoptysis prior to a fulminant haemorrhage or symptoms of uraemia, and some patients have no lung involvement at all and are simply investigated for haematuria or renal failure.

Patients often give a history of several months of non-specific symptoms with malaise, weight loss and headaches, but in contrast to patients with systemic vasculitis these are not prominent features. More commonly patients report minor haemoptysis over several weeks with symptoms of anaemia. The anaemia is usually microcytic with occasional features of microangiopathy, in contrast to the anaemia in systemic vasculitis which is usually normocytic. There is a strong association between pulmonary haemorrhage and cigarette smoking. The chest X-ray is rarely normal, but neither are the features specific for pulmonary haemorrhage. Alveolar shadowing usually involves the central lung fields, sparing the apices and supra-diaphragmatic regions. The opacities are often fleeting. These changes are very similar to those caused by pulmonary oedema or infection, however in haemorrhage they usually clear within 48 hours of an episode of bleeding. Other signs of pulmonary haemorrhage include sudden falls in haemoglobin and an elevated Kco - the diffusing capacity for carbon monoxide.

The range of renal presentation is similarly wide. At the severe end of the spectrum the progression of RPGN is inexorable with loss of renal function within days and little tendency to remit spontaneously. Very few untreated patients will retain useful renal function - only two of 52 patients in Benoit's 1964 review. The manifestations of the nephritis per se are indistinguishable from those of RPGN of any other cause with sudden onset, haematuria (usually microscopic), cellular casts and moderate proteinuria (less than 5g/24hrs). Hypertension is uncommon and oliguria is a late feature with a poor prognosis.

Aetiopathogenesis

All patients have circulating (serum) or bound anti-GBM autoantibodies which are clearly pathogenic. The target antigen for the autoantibodies is the non-collagenous domain of the 3 chain of type IV collagen, a basement membrane protein essentially restricted to the glomerular and alveolar basement membranes. Some patients produce antibodies against other basement membrane components, but these are not consistent findings. The precise antigenic epitopes have now been clarified using molecular dissection, modeled on the crystal structure of the non-collagenous domain. Anti GBM disease is strongly associated with specific major histocompatibility complex antigens (HLA alleles). Studies have shown that up to 90% of patients carry the HLA DR2 antigen, defined at a molecular level as the HLA B1*15 allele. HLA DR B1*01 and DRB1*07 seem to confer protection from the disease. T cells are almost certainly pivotal in the development of Goodpasture's disease, and recent experimental models have tentatively identified potential T cell epitopes bound on MHC molecules. Environmentally, exposure to hydrocarbons or solvent fumes has frequently been implicated in the development of Goodpasture’s disease, but it is difficult to establish causation.

Pathology

Light microscopic findings on renal biopsy vary considerably from entirely normal to severe necrotising crescentic nephritis, despite the fact that all cases display fixed immunoglobulin along the glomerular capillary walls in a linear pattern. It is common to see crescents in most glomeruli, all of a similar architecture (cellular or fibrous). An accompanying acute interstitial inflammatory cell infiltrate is common if not invariable. The blood vessels are usually normal. The characteristic linear staining of immunoglobulin along the GBM by direct immunofluoresence or immunohistochemistry is highly specific for Goodpasture's disease and occurs regardless of the serum antibody levels. Linear staining may rarely be seen in SLE, diabetes, myeloma and fibrillary glomerulonephritis (non-speciific binding).

Light microscopy of a single glomerulus from a patient with anti-GBM disease showing extensive crescent formation destroying Bowman’s capsule and fibrinoid necrosis	Immunofluorescence of a single glomerulus showing linear deposition of IgG along the glomerular basement membrane

Differential diagnosis

Other causes of RPGN must be clearly distinguished from anti-GBM disease, especially ANCA associated vasculitis. The blood serology and renal biopsy immunohistology will provide a clear distinction. It should be noted that a significant minority of patients carry both ANCA and anti-GBM antibodies.

Treatment

Treatment for almost all patients consists of plasma exchange, corticosteroids and cyclophosphamide. Without treatment most patient will die.

Treatment of anti-GBM disease

Initial treatment
Plasma exchange	Daily, 4 litre exchange for 5% human albumin solution. Use 300 – 600 mls fresh frozen plasma within 3 days of any invasive procedure (eg biopsy) or in patients with pulmonary haemorrhage. Continue for 14 days, or until antibody levels fully suppressed. Withhold if platelet count < 70 x 109/ml, or Hb < 9 g/dl. Watch for coagulopathy, hypocalcaemia and hypokalaemia.
Cyclophosphamide	Daily oral dosing at 3 mg/kg/day (round down to nearest 50 mg; reduce to 2 mg/kg/day in patients over 55 years). Stop if white cell count < 4 x 109/ml, and restart at lower dose when counts > 4 x 109/ml.
Prednisolone	Daily oral dosing at 1 mg/kg/day (maximum 60 mg). Reduce dose weekly to 20 mg by week 6, and then more slowly. No evidence for benefit of intravenous methylprednisolone and may increase infection risk (possibly use if plasma exchange not available).
Prophylactic treatments	Oral nystatin and amphotericin (or fluconazole) for oro-pharyngeal fungal infection. Ranitidine or proton-pump inhibitor for steroid promoted gastric ulceration. Low dose cotrimoxazole for Pneumocystis carinii pneumonia prevention. Consider aciclovir as cytomegalovirus prophylaxis. Consider calcium/vitamin D for prevention of osteoporosis (but relatively short course of steroids)
Maintenance treatment
Prednisolone	Reduce dose slowly from 20 mg at 6 weeks, to stop completely by 6 months.
Cyclophosphamide	Stop after 2-3 months. No further cytotoxic agents necessary.

Our most recent data suggests that response to treatment is strongly associated with extent of renal failure at time of presentation. Patients with serum creatinine less than 500 mcmol/L (5.6 mg/dL) at presentation will respond extremely well to intensive therapy, with 85% surviving 5 years, and most of these retaining independent renal function. In contrast, patients presenting oliguric or needing immediate dialysis respond poorly to treatment and only rarely recover renal function (8% of our patients in the long term). No patients with 100% crescents on renal biopsy and dialysis dependency recovered renal function. However, patients presenting with severe renal failure (creatinine > 500 mcmol/L) but not requiring immediate dialysis do respond in or hands to intensive therapy, 78% having independent renal function at 1 year, and 64% at last follow up or death.

	Number	Median follow-up (months)	5 year patient survival (%)	Median time to death (months)	Number of survivors with independent renal function at 5 years (%)	Number with independent renal function at death or follow-up (%)
Creatinine <500mol/l	20	108 (12-280)	17 (95%)	267 (12-280)	12 (86%)	15 (75%)
Creatinine >500 mol/l	11	98 (1-296)	7 (70%)	96 (1-120)	3 (43%)	7 (64%)
Dialysis dependent	38	109 (0.2-289)	14 (39%)	3 (0.2-237)	3 (21%)	3 (8%)
Total	69	104 (0.2-289)	39 (56%)	6 (0.2-280)	18 (52%)	25 (36%)

To achieve these results we institute 4L plasma exchange daily, initially for 14 days, using 5% human albumin solution as replacement fluid. This is continued until antibody levels are suppressed. Fresh frozen plasma is added in patients with pulmonary haemorrhage or within 3 days of renal biopsy or an invasive procedure. There is no evidence for any specific benefit of methyl prednisolone in Goodpasture's disease, but might be used as a temporising measure while waiting to transfer a patient to receive plasma exchange if not immediately available.

An alternative strategy is to remove circulating antibodies is by immunoadsorption with Staphylococcal protein A. The major advantage is the specific removal of immunoglobulin rather than depleting the patient of all circulating plasma proteins including clotting and complement factors, and avoiding the need for reinfusing large volumes of replacement blood products. However the technique is difficult to administer and not widely available.

After the acute phase of treatment, if antibody levels are suppressed and the patient stable, cyclophosphamide can be completely withdrawn at three months. Prednisolone is reduced gradually from 60mg to 20mg over the first 6 weeks, and then more slowly, aiming for cessation of treatment by 6 months. Autoantibody has usually disapeared by 8 weeks and rarely recurs. Long term immunosuppression is not needed. Close monitoring of the patient is essential during this period for signs of recrudesence of disease activity, iatrogenic haematological toxicity or infections. Regular measurement of circulating anti-GBM autoantibody is important to confirm adequacy of treatment and identify relapse. Serum creatinine is the best guide to renal function and should begin to decline within 1 week, whereas the urea usually rises with steroid therapy. Lung haemorrhage can be identified by changes in Kco, chest radiograph or haemoptysis, or by sudden falls in haemoglobin. The Kco is the most sensitive and specific investigation. Blood counts are crucial to detect haemorrhage and drug induced leucopenia. The white cell count rarely declines before 10 days, and classically falls two to three weeks after oral cyclophosphamide is begun. The dose of cyclophosphamide should then be reduced to avoid severe leucopenia.

Relapse and recurrence

Relapse is the worsening of signs and symptoms of the disease while circulating autoantibodies are still present. Unlike recurrence (reactivation of disease and autoantibody production after resolution of the auto-immune response) relapse is relatively common and may be precipitated by a number of insults. Local infections, fluid overload and cigarette smoking can precipitate fresh lung haemorrhage, while systemic infection can exacerbate nephritis and pulmonary bleeding, presumably by alterations in the cytokine milieu. Intravascular catheters and chest infections are the most important source of sepsis. Early discontinuation of plasmapheresis or immunosuppression may also allow disease relapse.

Recurrence is very unusual. Most patients in whom antibody production has abated remain free of any further episodes of either nephritis or pulmonary haemorrhage. However a few patients have recurrent mild pulmonary haemorrhage (usually precipitated by environmental toxins) with minimal nephritis, and rare patients present with distinct episodes of renal disease. Although such recurrences are almost always accompanied by a rise in antibody titre, detectable antibody production may also occur without apparent tissue injury. Renal transplantation can also precipitate recurrent autoantibody production if carried out before the auto-immune response has fully subsided.

References

Levy JB, Turner AN, Rees AJ & Pusey CD (2001). Long term outcome of anti-GBM disease treated by plasma exchange and immunosuppression. Annals Intern Med 134, 1033-42

Hudson BG, Tryggvason K, Sundaramoorthy M, Neilson EG (2003) Mechanisms of disease: Alport’s syndrome, Goodpasture’s syndrome and type IV collagen. 348; 2543-56

Wilson CB, Dixon FJ. Anti-glomerular basement membrane antibody induced glomerulonephritis. Kidney Int. 1973; 3: 74-89.

Lerner RA, Glassock RJ, Dixon FJ. The role of anti-glomerular basement membrane antibody in the pathogenesis of human glomerulonephritis. J Exp Med. 1967;126: 989-1004.

Levy JB, Lachmann RH, Pusey CD: Recurrent Goodpasture's Disease. Am J Kid Dis. 1996; 27: 573-8.

Johnson JP, Moore JJ, Austin HJ, Blalow JE, Antonovych TT, Wilson CB. Therapy of anti-glomerular basement membrane antibody disease: analysis of prognostic significance of clinical, pathological and treatment factors. Medicine. 1995; 64:219-27.

Levy JB, Pusey CD: Plasmapheresis, in Johnson R and Feehally J,eds Comprehensive Clinical Nephrology 2^nd Edition. Mosby: London, 2003