¹South Island Bone Marrow Transplant Unit, Christchurch Hospital, Christchurch, New Zealand

²Department of Endocrinology, Christchurch Hospital, Christchurch, New Zealand

Correspondence to: ProfD NJ Hart, Mater Medical Research Institute, Raymond Terrace, South Brisbane, Australia

cardiac failure is a known complication of haemopoietic stem cell transplantation (hsct) and is often difficult to diagnose as patients may have multiple medical problems. since brain natriuretic peptide (bnp) is largely a hormone of cardiac ventricular origin and is released early in the course of ventricular dysfunction, we have examined the value of serial plasma bnp levels for detecting cardiac failure in patients undergoing cytotoxic conditioning for hsct. fifteen patients undergoing hsct were evaluated (10 undergoing autologous hsct; five undergoing allogeneic hsct). bnp was measured by radioimmunoassay prior to therapy and weekly for 5 weeks. seven patients had a significant rise in bnp level (above a previously established threshold of 43 pmol/l associated with cardiac failure), occurring 1-4 weeks post commencement of conditioning. in three of these patients, cardiac failure was subsequently diagnosed clinically 3, 9 and 23 days after a bnp level of 43 pmol/l had been detected. these three patients had the highest peak bnp levels for the group and in each case elevation in bnp level occurred for a period exceeding 1 week. although numbers were relatively small, a bnp >43 pmol/l was significantly associated with the inclusion of high-dose cyclophosphamide in the preparative regimen (P = 0.02). BNP levels showed no relationship to febrile episodes. In conclusion, these results show that plasma BNP may be used as a marker for early detection of cardiac dysfunction in patients undergoing HSCT, particularly if levels are increased for periods exceeding 1 week. Measurement of BNP during HSCT may be helpful in patients at risk of cardiac failure, in complex clinical situations and in monitoring the cardiotoxicity of preparative regimens. Bone Marrow Transplantation (2000) 26, 309-313.

bone marrow transplantation; stem cell transplantation; brain natriuretic peptide; cardiac failure

The procedure of haemopoietic stem cell transplantation (HSCT) may injure the heart by a number of mechanisms.^1,2,3 Conditioning may incorporate cardiotoxic drugs, notably cyclophosphamide.^4,5 Sepsis is also common after HSCT and may produce focal and diffuse myocardial injury.⁶ Previous exposure to cardiotoxic drugs, notably anthracyclines, and transfusional iron overload may predispose patients to cardiac dysfunction. Heart failure may be precipitated by the use of hyperhydration regimens, transfused blood components, sodium retaining drugs and impaired renal function. The incidence of clinically significant cardiotoxicity has been estimated at between 5 and 10% of patients undergoing autologous or allogeneic HSCT.⁷

Early diagnosis of cardiac failure in patients undergoing HSCT may be difficult, especially in complex clinical situations where several pathologies often co-exist. For example, clinical and radiological signs may be obscured by lung infiltrates related to infection or graft-versus-host disease (GVHD); oedema and ascites are often due to other causes such as hypoalbuminaemia, veno-occlusive disease and capillary leak syndrome. Echocardiography is often helpful in such situations although it requires an experienced operator and may not be readily available in the acute situation. Cardiac catheterisation, the gold standard in the diagnosis of cardiac failure, is largely impractical due to its invasive nature and the attendant risk of bleeding or infection. Diagnosis of cardiac toxicity may be more relevant now that HSCT is being used for non-malignant diseases such as haemoglobinopathies, inborn errors of metabolism and autoimmune diseases which may in themselves be associated with underlying cardiac damage.

Brain natriuretic peptide (BNP) was originally discovered in the porcine brain⁸ but was subsequently found to be predominantly a cardiac hormone.⁹ Unlike atrial natriuretic peptide (ANP), which is secreted by the atria in response to increased atrial pressure¹⁰ BNP is derived chiefly from the cardiac ventricles in response to ventricular stresses.¹¹ The plasma levels of both peptides are inversely correlated with measures of cardiac function⁹ and recent work has shown BNP to be a more sensitive marker of cardiac impairment than ANP.^12,13 Indeed, plasma BNP levels are useful in differentiating acute heart failure from pulmonary disease¹⁴ and for the early diagnosis of heart failure.¹⁵ The aims of this study were to investigate plasma BNP as a marker of ventricular dysfunction in patients receiving high-dose preparative regimens and HSCT.

Patients and methods

Patients

The study was approved by the Ethics Committee of Southern Regional Health Authority (Canterbury, New Zealand) and all patients gave informed consent. Patients received a variety of preparative regimens including fractionated total body irradiation (TBI, 12 Gy given as 133 cGy three times daily over 3 days) in combination with cyclophosphamide 120 mg/kg, etoposide 60 mg/kg¹⁶ or melphalan 140 mg/kg,¹⁷ BEAM (BCNU 300 mg/m², etoposide 800 mg/m², ara-C, 1600 mg/m², melphalan 140 mg/m²)¹⁸ CBV (cyclo- phosphamide 7.2 g/m², BCNU 300 mg/m², etoposide 600 mg/m²)¹⁹ single agent busulphan 16 mg/kg, and high-dose carboplatin (dose adjusted for glomerular filtration rate and surface area) combined with melphalan 180 mg/m². All patients receiving high-dose cyclophosphamide received intravenous mesna. Patients undergoing allogeneic HSCT received GVHD prophylaxis in the form of short methotrexate and cyclosporin A^20,21 and treatment of acute GVHD followed previously published protocols.²²

All patients were managed according to the South Island Bone Marrow Transplant Unit (SIBMTU) Protocol for Haemopoietic Stem Cell Transplantation which included full daily clinical examination, 4-6 h temperature assessment, daily weight, daily full blood count and biochemistry, prophylactic antibiotics, transfusion of red cells to maintain haemoglobin >80 g/l and of platelets to maintain a platelet count of >10 ´ 10⁹/l, if condition stable, and 20 ´ 10⁹/l, if condition unstable (eg febrile). Intravenous fluids were given to balance the deficit between oral intake and output (including insensible losses). If fluid excess was noted clinically, or from weight gain or fluid balance recordings then correction would be carried out by reducing intake plus the cautious use of small amounts of frusemide (20-40 mg intravenously). ECGs were routinely performed prior to commencing conditioning and chest X-rays were performed weekly throughout the neutropenic period and as clinically indicated.

The following parameters were also documented for each patient: left ventricular ejection fraction (LVEF) by gated radionucleide ventriculography scan, pre-HSCT and at 6-8 weeks after HSCT, pre-HSCT ferritin (as an estimate of iron loading), and history of treatment with anthracyclines. Cardiac failure was diagnosed as a clinical syndrome consisting of appropriate symptoms (shortness of breath, fluid retention, fatigue) and clinical signs of fluid retention (pulmonary or peripheral) supported by chest radiographic appearances or reduced LVEF <50% or a beneficial response to therapy for heart failure. The clinical diagnosis of heart failure, was made independently of any knowledge of the BNP results.

Assay of BNP

EDTA plasma (5 ml) was collected pre-transplant and then weekly. Samples were centrifuged within 1 h, separated and the plasma stored at -80°C before assay. Plasma aliquots (1 ml) were extracted on Sep Pak cartridges pre-equilibrated with methanol, washed with 0.1% trifluoroacetic acid and eluted with 2 ml 80% propranolol:0.1% trifluoroacetic acid (80:20 v/v). After pre-incubation for 3 h at room temperature, trace was added and the assay incubated for 18 h at 4°C. The rest of the assay was as previously described,²³ as modified for clinical use.²⁴ Coefficient of variation at 35 pmol/l was 6% and 11% for intra- and inter-assay, respectively. A BNP level of 43 pmol/l was considered consistent with the diagnosis of heart failure.²⁴

Statistical analysis

Data were analysed using Mann-Whitney U test for comparison of grouped continuous variables, chi-squared test for grouped discrete variables and Spearman's test for correlation. The data were dichotomised into patients with a BNP level of greater or less than 43 pmol/l.²⁴

Results

Patient data

Patient data are summarised in Table 1. All patients had normal ECGs and no clinical evidence of cardiac failure prior to HSCT. Three patients (patients 4, 6 and 11, Table 1) developed clinical signs of cardiac failure during HSCT. In patient 4, the diagnosis was made on clinical grounds and response to diuretic therapy, in patient 6 the diagnosis was made on clinical grounds and a simultaneous LVEF of 49% (falling from 66% pre-transplant) and in patient 11 the diagnosis was made on clinical and radiological grounds. Patients 6 and 11 also responded to diuretic therapy.

BNP levels during HSCT

The changes in plasma BNP level over the 5 weeks following the commencement of conditioning are summarized in Figure 1. In seven patients the threshold value of BNP (43 pmol/l) was exceeded at least once during the study period, BNP peaking between 1 and 4 weeks (median 3 weeks) post high-dose therapy. In the three patients (patients 4, 6 and 11) in whom a clinical diagnosis of cardiac failure was made, a rise in the BNP level above 43 pmol/l occurred 23, 9 and 3 days, respectively, before the clinical signs became apparent. These three patients had the highest peak BNP levels of the group and, in addition, plasma BNP levels were elevated for at least 1 week in all cases. The remaining four patients had lesser elevations of BNP on only single occasions and showed no clinical or radiological signs of cardiac failure. There were no patients in whom clinical cardiac failure occurred without an associated rise in the BNP level above the 43 pmol/l threshold.

Levels of BNP showed no correlation with fever defined either as a mean of the temperatures taken on the same day as testing (r = 0.001, P = 0.9) or as the occurrence of a temperature of >38°C in the previous week (P = 0.6). However, the BNP level was correlated with simultaneous plasma creatinine level (r = 0.592, P < 0.0005). Except for patient 11 whose peak 'simultaneous' creatinine level was 0.22 mol/l, all other creatinines were within the normal range and did not exceed 0.11 mol/l.

Dichotomisation of data using the threshold value of 43 pmol/l

Table 2 provides a comparison of the two groups dichotomised by the BNP level of 43 pmol/l. In both groups, a similar proportion of patients underwent autologous and allogeneic HSCT and there were no significant differences in baseline LVEF or incidence of previous anthracycline exposure between the two groups. There was a significant difference in the number of patients receiving cyclophosphamide between the two groups (P = 0.02), with a greater number of those with raised BNP levels (>43 pmol/l) receiving cyclophosphamide-containing regimens. There was also a trend towards greater age (P = 0.07) and higher ferritin levels (P = 0.11) in this group, although these did not reach significance.

Discussion

This study supports the use of plasma BNP as a marker of ventricular dysfunction in patients undergoing high-dose therapy and HSCT. Raised plasma BNP levels have previously been shown to be early markers of cardiac impairment, heralding the clinical picture by days to weeks^25,26 and are generally superior to ANP as markers of ventricular dysfunction.²⁷ In the present study, those subjects who developed clinical heart failure exhibited raised plasma BNP for periods exceeding 1 week prior to the diagnosis being made. A raised BNP level, although not diagnostic alone, might therefore be used in the context of HSCT to assist diagnosis when a complex clinical situation is encountered, to exclude cardiac failure or to select patients for echocardiography. Febrile episodes do not appear to confound the interpretation of raised BNP levels, which is reassuring since fever is common following HSCT and can result in impaired cardiac function.²⁸ The significance of the association of BNP level with creatinine level in these patients is uncertain, as all patients except for patient 6 had creatinine levels in the normal range. One reason may be the close relationship between cardiac and renal function in relation to fluid balance, together with a confounding effect of age which is associated with increases in both plasma creatinine and BNP levels.²³ Although plasma BNP increases with severe chronic renal failure (possibly because of associated heart failure), there is no relationship in subjects with plasma creatinine less than 0.2 mmol/l. Certainly, plasma creatinine levels are not predictive for cardiovascular events in patients suffering myocardial damage from infarction, in strong contrast to plasma BNP levels.^27,29

The reasons for the raised plasma BNP level in the four patients who were not considered to have cardiac failure are not obvious, but may represent a more transient period of cardiac damage insufficient to cause cardiac decompensation. Raising the threshold level of BNP might result in greater specificity in this group of patients. This is supported by patients 4, 6 and 11 who had the highest peak BNP levels. However, all of these patients had raised plasma BNP levels for time periods exceeding 1 week well prior to the onset of clinical symptoms and signs of heart failure. A sustained rise in plasma BNP may therefore be a better discriminator for the diagnosis of cardiac failure than an increased threshold level which could reduce the sensitivity for early detection of cardiac failure in such patients. It should be noted that in subjects with overt cardiac failure, plasma BNP is clearly elevated, in contrast to LVEF which may be variably affected, and within the reference range in an important minority of patients.^27,29 Some of these patients primarily have diastolic dysfunction.¹⁴ Taken together, it is not surprising that plasma BNP after a cardiac insult will be increased, and not necessarily reflected by a persistent fall in LVEF, as demonstrated in this study.

The toxic effects of cyclophosphamide on the myocardium have been documented in HSCT patients using other methods including serial ECGs,⁷ echocardiography,³⁰ PET scanning³¹ and, in severe cases, autopsy findings of haemorrhagic myocardial necrosis.^4,30,32 Data suggest that dosing may be an important risk factor.⁵ Echocardiographic abnormalities develop in over half of paediatric patients 1 week after and persist for several weeks beyond high-dose cyclophosphamide administration.¹ Our findings using raised BNP levels as markers of cardiac impairment are similar, and using BNP as a marker, it is notable that the maximal effect in most patients is not immediate but delayed for 1-4 weeks. Clinicians should therefore be aware that susceptibility to cardiac failure is not restricted to timing of high-dose therapy and hyperhydration, but represents an ongoing risk for several weeks. Although animal data suggest that mesna may prevent cyclophosphamide cardiotoxicity,¹ this measure (taken to prevent haemorrhagic cystitis) did not prevent a significant rise in plasma BNP in most patients receiving cyclophosphamide in this study.

Previous exposure to anthracyclines may have been important in the development of cardiac dysfunction and one previous study has used ANP as a marker in patients receiving anthracycline-containing regimens.³³ Analysis of previous anthracycline exposure as a risk factor was not possible as numbers were insufficient to examine varying degrees of exposure to different anthracyclines. Although irradiation is known to cause cardiac damage in higher doses, for example, such as those used previously in mantle irradiation for Hodgkin's disease,³⁴ it is presently unclear whether the lower, fractionated doses of TBI used in HSCT are associated with acute or chronic cardiotoxicity. In the present study, most patients received TBI as part of their preparative regimen and there was no apparent relationship with subsequent BNP levels, although numbers were small. Age and transfusional iron overload may also have contributed to ventricular dysfunction during HSCT. There was a trend for elevated ferritin levels in patients with raised BNP levels although statistically these effects were not significant, possibly due to the small sample size. Analysis of larger numbers of patients may yield more information regarding how multiple risk factors, including different conditioning regimens, act together to produce ventricular dysfunction.

Elevated plasma BNP levels are a simple, sensitive and specific predictor of impending cardiac dysfunction after HSCT and as such may be a useful adjunct to current methods of patient monitoring in this setting. In addition, the use of this assay in prospective trials will facilitate the design of conditioning regimens that minimise cardiac toxicity.

Acknowledgements

We would like to acknowledge the SIBMTU team for their care of these patients, Dr Tim Yandle, Steve Fisher and Sarah Raudsepp for BNP assay, and Mr Ian Nivison-Smith, Australian Bone Marrow Transplant Recipient Registry, for statistical advice. This study was supported in part by grants from the New Zealand Health Research Council and the National Heart Foundation of New Zealand.

1 Petersen FB, Bearman SI. Preparative regimens and their toxicity. In: Forman SJ, Blume KG, Thomas ED (eds). Bone Marrow Transplantation. Blackwell Scientific Publications: Boston, 1994, pp82-83.

2 Feneley MP, Lim CA. Cardiac complications. In: Atkinson K (ed). Clinical Bone Marrow and Stem Cell Transplantation. Cambridge University Press: Cambridge, 2000, pp952-957.

3 Cazin B, Gorin NC, Laporte JP et al. Cardiac complications after bone marrow transplantation. A report on 63 consecutive transplantations. Cancer 1986; 57: 2061-2069, MEDLINE

4 Gottdiener JS, Appelbaum FR, Ferrans VJ et al. Cardiotoxicity associated with high dose cyclophosphamide therapy. Arch Intern Med 1981; 141: 758-763, MEDLINE

5 Goldberg M, Antin JH, Guinin EC et al. Cyclophosphamide cardiotoxicity: an analysis of dosing as a risk factor. Blood 1986; 68: 1114-1118, MEDLINE

6 Parillo JE. Cardiovascular dysfunction in septic shock: new insights into a deadly disease. Int J Cardiol 1985; 7: 314-321, MEDLINE

7 Kupari M, Volin L, Suokas A et al. Cardiac involvement in bone marrow transplantation: electrocardiographic changes, arrhythmia, heart failure and autopsy findings. Bone Marrow Transplant 1990; 5: 91-98, MEDLINE

8 Sudoh T, Kangawa K, Minamino N et al. A new natriuretic peptide in porcine brain. Nature 1988; 332: 78-81, MEDLINE

9 Mukoyama M, Nakao K, Hosoda K et al. Brain natriuretic peptide as a novel cardiac hormone in humans. J Clin Invest 1991; 87: 1402-1412, MEDLINE

10 Lang RE, Tholken H, Ganten D et al. Atrial natriuretic factor - a circulating hormone stimulated by volume loading. Nature 1985; 314: 264-267, MEDLINE

11 Yasue H, Yoshimura M, Sumida H et al. Localisation and mechanism of secretion of B-type natriuretic peptide in normal subjects and patients with heart failure. Circulation 1994; 90: 195-200, MEDLINE

12 Choy AJ, Darber D, Lang C et al. Detection of left ventricular dysfunction after acute myocardial infarction: comparison of clinical, echocardiographic, and neurohormonal methods. Br Heart J 1994; 72: 16-22, MEDLINE

13 Davidson NC, Naas AA, Hanson JK et al. Comparison ofatrial natriuretic peptide, B-type natriuretic peptide, and N-terminal proatrial natriuretic peptide as indicators of leftventricular systolic dysfunction. Am J Cardiol 1996; 77: 828-831, MEDLINE

14 Davis M, Espiner E, Richards G et al. Plasma BNP in assessment of acute dysnoea. Lancet 1994; 343: 440-443, MEDLINE

15 Cowie MR, Struthers AD, Wood DA et al. Value of natriuretic peptides in the assessment of patients with possible new heart failure in primary care. Lancet 1997; 350: 1347-1351,

16 Blume K, Forman S, O'Donnell MR et al. Total body irradiation and high dose etoposide: a new preparatory regimen for bone marrow transplant in patients with advanced hematologic malignancies. Blood 1987; 69: 1015-1020, MEDLINE

17 Powles R, Milliken S, Helenglass G et al. The use of melphalan in conjunction with total body irradiation as treatment for acute leukaemia. Transplant Proc 1989; 21: 2955-2957, MEDLINE

18 Linch DC, Winfield D, Goldstone AH et al. Dose intensification with autologous bone-marrow transplantation in relapsed and resistent Hodgkin's disease: results of a BNLI randomised trial. Lancet 1993; 341: 1051-1054, MEDLINE

19 Jagannath S, Dicke KA, Armitage JO et al. High-dose cyclophosphamide, carmustine, and etoposide and autologous bone marrow transplantation for patients with advanced Hodgkin's disease. Ann Intern Med 1986; 104: 163-168, MEDLINE

20 Storb R, Deeg HJ, Whitehead J et al. Methotrexate and cyclosporine compared with cyclosporine alone for prophylaxis of acute graft-versus-host disease after marrow transplant for leukemia. New Engl J Med 1986; 314: 729-735, MEDLINE

21 Storb R, Deeg HJ, Pepe M et al. Methotrexate and cyclosporine versus cyclosporine alone for prophylaxis of graft-versus-host disease in patients given HLA-identical marrow grafts for leukemia. Blood 1989; 73: 1729-1734, MEDLINE

22 Sullivan K. Graft-versus-host disease. In: Forman SJ, Blume KG, Thomas ED. Bone Marrow Transplantation. Blackwell Scientific Publications: Boston, 1994, pp339-362.

23 Yandle TG, Richards AM, Gilbert A et al. Assay of brain natriuretic peptide (BNP) in human plasma: evidence for high molecular weight BNP as a major plasma component in heart failure. J Clin Endocrinol Metab 1993; 76: 832-838, MEDLINE

24 Fleischer D, Espiner EA, Yandle TG et al. Rapid assay of plasma brain natriuretic peptide in the assessment of acute dyspnoea. NZ Med J 1996; 100: 71-74,

25 Richards AM, Crozier IG, Yandle TG et al. Brain natriuretic factor: regional plasma concentrations and correlations with haematodynamic state in cardiac disease. Br Heart J 1993; 69: 414-417, MEDLINE

26 Hunt PJ, Richards AM, Nicholls MG et al. Immunoreactive amino-terminal pro-brain natriuretic peptide (NT-proBNP): a new marker of cardiac disease. Clin Endocrinol 1997; 47: 287-296,

27 Richards AM, Nicholls MG, Yandle TG et al. Plasma N-terminal pro-brain natriuretic and adrenomedulin. New neurohormonal predictors of left ventricular function and prognosis after myocardial infarction. Circulation 1998; 97: 1921-1929, MEDLINE

28 Haupt MT, Rackow ED. Adverse effects of febrile state on cardiac performance. Am Heart J 1983; 105: 763-768, MEDLINE

29 Richards AM, Nicholls MG, Yandle TG et al. Neuroendocrine prediction of left ventricular function and heart failure after acute myocardial infarction. Heart 1999; 81: 114-120, MEDLINE

30 Kupari M, Volin L, Suokas A et al. Cardiac involvement in bone marrow transplantation: serial changes in left ventricular size, mass and performance. J Intern Med 1990; 227: 259-266, MEDLINE

31 Gardner SF, Lazarus HM, Bednarczyk EM et al. High-dose cyclophosphamide-induced myocardial damage during BMT: assessment by positron emission tomography. Bone Marrow Transplant 1993; 12: 139-144, MEDLINE

32 Buja LM, Ferrans VJ, Graw RG. Cardiac pathologic findings in patients treated with bone marrow transplantation. Hum Pathol 1976; 7: 17-45, MEDLINE

33 Bauch M, Ester A, Kimura B et al. Atrial natriuretic peptide as a market for doxorubicin-induced cardiotoxic effects. Cancer 1992; 69: 1492-1497, MEDLINE

34 Aisenburg AC. Problems in Hodgkin's disease management. Blood 1999; 93: 761-779, MEDLINE

Figure 1 Changes in plasma BNP levels during HSCT in 15 patients. 43 pmol/l denotes a previously established threshold associated with cardiac failure.²⁴ Patients exceeding this threshold are indicated. Patients 4, 6 and 11, who had the highest peak levels of BNP and raised levels exceeding 1 week, developed clinical symptoms of cardiac failure. Patients whose BNP level remained below the threshold of 43 pmol/l are represented by broken lines.

Table 1  Patient data

Table 2  Analysis of groups dichotomised by peak BNP levels of above or below 43 pmol/l