Original Article

Mortality Patterns in Childhood Lupus – 10 Years’ Experience in aDeveloping Country

S. Singh, Devidayal, L. Kumar and K. Joshi

Postgraduate Institute of Medical Education and Research, Chandigarh, India

Abstract: Over the last 10 years we have seen 31children with systemic lupus: 10 (32%) of these havedied. The commonest primary determinant of mortalitywas uncontrolled disease activity (60%), possiblycontributed to by late referrals leading to delays indiagnosis and the institution of therapy. Thromboem-bolic disease was responsible for two deaths andtuberculosis for one. The major contributory factorswere nephritis in two patients and myocarditis, en-docarditis, lupus pneumonia and CNS disease in onepatient each. Concomitant infections, predominantlynosocomial, occurred in a significant proportion ofpatients (40%) but were only cofactors in mortality.Autopsies were done in three cases. These patterns ofdeath are significantly different from those seen in thedeveloped world, where disease activity has ceased to bean important factor owing to early recognition of cases.Avoiding late referrals and delays in diagnosis canreduce mortality in childhood-onset lupus.

Keywords: Childhood lupus; Developing countries;Mortality patterns; Uncontrolled disease activity

Introduction

The prognosis of childhood-onset systemic lupuserythematosus (SLE) has significantly improved overthe last few decades. Despite the improved outlook,

however, a significant proportion of children stillsuccumb to the illness. Although there is good amountof literature on the causes of death in childhood-onsetSLE from the developed world, the data from developingcountries (specially autopsy data) are virtually non-existent. We examined the mortality patterns in aninception cohort of 31 children with SLE diagnosed andfollowed up in the Pediatric Rheumatology andImmunology Clinic of the Post Graduate Institute ofMedical Education and Research, Chandigarh.

Patients and Methods

The study cohort consisted of 31 children who weretreated for SLE between February 1991 and February2001. The diagnosis was based on the revised 1982American College of Rheumatology criteria [1]. Amedian of six (range 4–8) ACR criteria was fulfilledby all patients. Disease activity was defined in terms of aflare in one or more organ systems requiring an increasein suppressive therapy, a persistent moderate to high-grade fever of undetermined aetiology, and/or elevatedsedimentation rates. Ten out of these 31 patients (32%)died during a mean follow-up of 3.2 years (range 5months to 10 years). A review of the medical records ofthese 10 children, along with postmortem findings inthree of them, was performed with the aim ofascertaining the cause of death (Table 1).

Their ages ranged from 6.5 to 12 years (mean 10.6);there were eight girls and two boys. The duration ofdisease varied from 20 days to 2 years (mean 9.4months). Duration of activity at presentation averaged82.3 days (range 14–150) in those who ultimately died ofactive lupus; this was longer by 46.8 days than in thosein the original cohort who survived disease activity

Clin Rheumatol (2002) 21:462–465� 2002 Clinical Rheumatology Clinical

Rheumatology

Correspondence and offprint requests to: Dr Surjit Singh, AdditionalProfessor of Pediatric Allergy and Immunology, Advanced PediatricCentre, Post Graduate Institute of Medical Education and Research,Chandigarh-160012, India. Tel: 0091-172-747585 ex. 504; 235 (Off.);0091-172-710786 (Res.); Fax: 0091-172-744401; 0091-172-745078;E-mail: surjitsingh@i91.net.in; apcpgi@glide.net.in

(average 35.5 days, range 7–120). Six out of 10 patientsdied within an average 20.3 days of diagnosis (range 7–30 days); two died within a year and two within 2 yearsof presentation (mean follow-up 18.5 months). Theproportionate mortality was similar in both sexes(female, 8/25=32%; male, 2/6=33.3%).

All patients had received corticosteroids (usuallyprednisolone 1–2 mg/kg/day) for control of diseaseactivity. Five patients who died during the initialadmission itself were given intravenous steroids aspulses (either methylprednisolone 30 mg/kg/day ordexamethasone 5 mg/kg/day). Patients with evidence

of significant renal involvement also received cyclo-phosphamide pulse therapy in addition. Other thera-peutic modalities, such as antitubercular therapy,antihypertensive agents, anticoagulants, intravenousimmunoglobulins, peritoneal dialysis and ventilation,were used according to the clinical indications [2].

All 10 patients who died had had multisysteminvolvement, and multiple factors seemed to beresponsible for their mortality as well, with theproximate cause of death being identifiable in majority.One patient (no. 1), however, had a sudden unexplaineddeath at another hospital.

Table 1. Clinical details of 10 patients of SLE included in the study

S.no.

Age (yrs)and sex

Major systemsinvolved

Clinicalcourse

Main therapiesused

Main causeof death

Majorcontributoryfactors

Timesincediagnosis

Significant autopsyfindings

1. 8/F SkinJointsGI

Remissionon steroids

HydrocortisonePrednisoloneChloroquine

Unknown ?fulminantsepsis

2 years Died elsewhere

2. 11/F SkinJointsCVSCNSRenal

Died in firstadmission

Pulse MPand CPVentilation

UDA MyocarditisNephritis

10 days MyocarditisSerositisHypoxic changesand brain oedemaClass IV nephritis

3. 8/M SkinGICNSRenal

Died in firstadmission

Pulse dexaAntibiotics

UDA EndocarditisCNS diseaseNephritis(class IV)

1 month Not done

4. 9/F SkinGIJointsPulmonary

Died in firstadmission

Pulse dexaand CPVentilation

UDA Nephritis(class IV)Myocarditis

2 weeks Not done

5. 10/F SkinJointsPulmonaryCNSRenal

Died in firstadmission

Pulse dexaand CPVentilationIVIG

UDA Lupuspneumonia,S. aureussepsis,Nephritis (IV)

1 month Not done

6. 10/F SkinJointsCNSRenalCVS

Arthralgiaspersisted,died after2nd flare

Pulse CPPrednisoloneATT

UDA CNS diseaseSepsisTuberculosisNephritis(class III)

8 months Tubercularabscesses inspleenHilar nodetuberculosis

7. 10/M SkinLiverJointsKidneysPulmonary

Remissionon steroidsShuntmalfunction

PrednisolonePulse dexaATTVP shunt

Tuberculosis Shuntmalfunction

1 year Not done

8. 6.5/F SkinJointsLiver, CNSRenal

Died in firstadmission

Pulse MPPeritonealdialysis,Ventilation

UDA Nephritis(Class IV)CNS diseaseMyocarditis

1 week Not done

9. 12/F SkinJointsCVSRenalPulmonary

Recurrentflares requiringhospitalisation(7 times)

PrednisoloneATTPulse dexaHeparin

Systemicthromboembolism

Nephritis(Class V)

2.5 years Massivepulmonaryembolism (acute)RVT (organized)Nephritis Class VbPAH (grade III)Infarct basalganglia

10. 10/F PulmonarySkin

Died after firstadmission

PrednisoloneAspirin

Severe PAH(micro-thromboembolism)

– 1 month Died at home

Abbreviations: MP, methylprednisolone; CP, cyclophosphamide; dexa, dexamethasone; IVIG, intravenous immunoglobulins; ATT, antituberculartherapy; UDA, uncontrolled disease activity; PAH, pulmonary artery hypertension; RVT, renal vein thrombosis.

Mortality in Childhood Lupus 463

Uncontrolled disease activity was considered to be themain determinant of mortality in six patients. Five ofthese (nos. 2–5 and 8) had died during the first admissionfrom severe disease refractory to even pulse therapy withintravenous steroids and cyclophosphamide. The clinicalcourse of the other patient (no. 6) who died of activelupus during follow-up, was marked by disease flaresand she died during her second flare after diagnosis.Thromboembolism due to antiphospholipid syndrome

was the second commonest primary cause of mortality.Patient no. 9 (reported elsewhere) [3] had severepulmonary arterial hypertension (diagnosed on cardiaccatheterisation), possibly resulting from recurrent smallthromboembolic episodes related to lupus anticoagulantactivity. She did not respond to therapy and was reportedto have died at home subsequently. The other patientwho died apparently of pulmonary thromboembolismhad evidence of a more widespread thromboembolicdisease on autopsy. Lupus anticoagulant was positive inboth these cases.Nephritis (WHO class III–V) was seen to occur in

seven patients and manifested with either proteinuriaalone (no.7) or in combination with haematuria,azotaemia and hypertension. In two patients (nos. 4and 8) nephritis resulted in acute renal failure.Myocarditis was a significant factor for death in onepatient (no. 2). Echocardiography in this patient showedseverely impaired left ventricular function with anejection fraction of 18%; subsequent autopsy confirmedmyocardial involvement [4]. Acinetobacter endocarditiswas a major contributory factor in one patient (no.3).Large vegetations were seen on the mitral valve onechocardiography, and the sudden demise was possiblydue to massive myocardial infarction or embolism.Autopsy could not be done in this case. Patient no. 6 hadpredominantly neurological involvement resulting inrefractory status epilepticus. Her CT scan had shownbilateral cortical atrophy with ex vacuo hydrocephalus.She also had Enterobacter and Salmonella typhimuriumsepsis and evidence of disseminated tuberculosis onautopsy. Lupus pneumonia made a major contribution todeath in one patient (no. 5). Chest X-ray showedbilateral whiteout lung fields, which progressed toacute respiratory distress syndrome. Echocardiographyin this patient was normal.Concomitant infections were seen in four patients.

Bacterial isolates (Enterobacter, Salmonella typhimur-ium, Staphylococcus aureus and Acinetobacter) in threepatients were thought to be hospital acquired. Tubercularinfection was proved in two patients; this was the solefactor causing death in one (no.7), who had developedtubercular meningitis on follow-up, leading to hydro-cephalus, underwent revision surgery for shunt malfunc-tion and subsequently died of uncontrolled raisedintracranial pressure.Although the pattern of organ involvement was

similar in survivors and non-survivors, the latterappeared to have multiorgan involvement right fromthe beginning. Severe nephritis (class III and IV) was

also more common in non-survivors (5/7=71% vs 4/13=30%).

Discussion

Cook et al. [5], in 1960, described childhood SLE as‘generally a progressive disease terminating fatally’.Since that time there has been considerable improvementin the prognosis for this condition. The survival rates,which were as low as 44% at 2 years in the early 1960s,have now improved to 85%–95% at 10 years [6–9].These figures reflect the situation in the developedworld, where SLE does not now appear to materiallyaffect the long-term survival of patients [10]. Improve-ments in prognosis have been related to early identifica-tion of cases [11], the use of potent anti-inflammatoryand immunosuppressive drugs [12], an efficient health-care system [10] and continuity of follow-up care [13].

In developing countries the poorer prognosis is due toa multitude of factors, especially poor socioeconomicand educational status causing delay in seeking medicaladvice, long travelling distances to reach hospitalscapable of caring for such patients, misdiagnosis byprimary care physicians thus delaying appropriatetreatment, and late referrals to a tertiary care center,with the patient in a critical state when disease activitybecomes difficult to control [14]. Some of these factorsseemed to operate in our patients also, particularly thosedying early. Children who died soon after initialhospitalisation had been referred in a critical state anddied of uncontrolled disease activity, despite aggressivetherapy that included intravenous corticosteroids andcyclophosphamide. These findings are in agreement withthe reports of Meislin et al. and Cassidy et al., who alsofound that early phase deaths were usually due to lupusactivity [15,16]. Early recognition of cases has virtuallyeliminated active disease as a predominant cause ofdeath in developed countries now [17], but this is stillnot the case in developing countries such as our own.Late referrals have therefore contributed significantly tothe relatively poor survival rates in our patientscompared to those reported in recent series [9,18,19].However it is also possible that some of these childrenmay have had refractory lupus, which is sometimesunresponsive to all forms of therapy.

We found that thromboembolic phenomena weredirectly responsible for two deaths in the presentseries. This has been infrequently related to mortalityin childhood lupus [8,20]. Nephritis is an importantcause of death in childhood SLE [16,18]. Althoughseven out of 10 patients in the present series hadsignificant renal involvement on histology, renal failureper se did not appear to be an important primary cause ofdeath in our patients.

Infection, a reflection of basic immune defectsinherent in the disease process and the superimposedeffects of corticosteroids and immunosuppressive ther-apy, has been identified as a major factor in mortality inSLE [17–19,21]. In our patients, although bacterial

464 S. Singh et al.

infections could be taken as cofactors in mortality inhalf, and mycobacterial tuberculosis being directlyresponsible for one death, the majority of such infectionswere nosocomial and occurred in a setting of uncon-trolled disease activity, possibly as part of terminalevents in what would even otherwise have been fatalSLE.

Our experience shows that the first 4 weeks afterdiagnosis are the most crucial as far as survival inchildhood SLE is concerned. Early mortality can besignificantly decreased if delays in diagnosis can beavoided and child is referred to an appropriate centrewhere expertise in the management of this condition isavailable. Meticulous follow-up significantly improvessurvival [13].

References

1. Tan EM, Cohen AS, Fries JF et al. The 1982 revised criteria forthe classification of systemic lupus erythematosus. ArthritisRheum 1982;25:1271–7.

2. Singh S, Kumar L, Khetarpal R et al. Clinical and immunologicalprofile of SLE: some unusual features. Indian Pediatr1995;34:979–86.

3. Khetarpal R, Goraya JS, Singh S, Singh M, Kumar L. Pulmonaryhypertension as presenting feature of childhood SLE: associationwith lupus anticoagulant. Scand J Rheumatol 1997;26:325–6.

4. Singh S, Radotra B. Clinicopathological conference – a child withpainful joints and failing kidneys. Indian Pediatr 1996;33:932–43.

5. Cook CD, Wedgwood RJP, Craig JM, Hartmann JR, JanewayCA. Systemic lupus erythematosus. Description of 37 cases inchildren and a discussion of endocrine therapy in 32 of the cases.Pediatrics 1960;26:570–85.

6. Abeles M, Urman JD, Weinstein A et al. Systemic lupuserythematosus in the younger patient: survival studies. JRheumatol 1980;7:515–22.

7. Yancey CL, Doughty RA, Athreya BH. Central nervous systeminvolvement in childhood systemic lupus erythematosus. ArthritisRheum 1981;24:1389–95.

8. Glidden RS, Mantzouranis EC, Borel Y. Systemic lupuserythematosus in childhood: clinical manifestations and improvedsurvival in fifty-five patients. Clin Immunol Immunopathol1983;29:196–210.

9. Bardare M, De Vio M, Giani M, Cohen E. Systemic lupuserythematosus in childhood: review of literature and personalobservations on 32 cases. Pediatr Med Chir 1990;12:577–86.

10. Jonsson H, Nived O, Sturfelt G. Outcome in systemic lupuserythematosus: a prospective study of patients from a definedpopulation. Medicine 1989;68:141–50.

11. Grigor RR, Edmonds J, Lewkonia R, Bresnihan B, Hughes GRV.Systemic lupus erythematosus: a prospective analysis. AnnRheum Dis 1978;37:121–8.

12. Appel GB, Cohen DG, Pirani CL, Meltzer JI, Estes D. Long-termfollow-up of patients with lupus nephritis. A study based onclassification of World Health Organisation. Am J Med1987;83:877–85.

13. Wallace DJ, Podell T, Weiner J, Klinenberg JR, Forouzesh S,Dubois EL. Systemic lupus erythematosus – survival patterns.JAMA 1981;245:934–8.

14. Kumar A, Malaviya AN, Singh RR, Singh YN, Adya CM, KakkarR. Survival in patients with systemic lupus erythematosus inIndia. Rheumatol Int 1992;12:107–9.

15. Meislin AG, Rothfield N. Systemic lupus erythematosus inchildhood. Pediatrics 1968;42: 37–49.

16. Cassidy JT, Sullivan DB, Petty RE, Ragsdale C. Lupus nephritisand encephalopathy. Prognosis in 58 children. Arthritis Rheum1977;20(Suppl 2):315–22.

17. Levy M, Montes de Oca M, Claude-Babron M. Unfavourableoutcomes (end-stage renal failure/death) in childhood onsetsystemic lupus erythematosus. A multicenter study in Paris andits environs. Clin Exp Rheumatol 1994;12(Suppl):S63–8.

18. Lee BW, Yap HK, Yip YC. A 10 year review of systemic lupuserythematosus in Singapore children. Aust Pediatr J1987;23:163–5.

19. Huang JL, Lin CJ, Hung IC, Luo SF. The morbidity and mortalityassociated with childhood onset systemic lupus erythematosus.Chang Keng I Hsuch Tsa Chih 1994;17:113–20.

20. Chandrashekran AN, Rajenderan CP, Ramakrishnan S, MadhavanR, Pratibhan M. Childhood systemic lupus erythematosus inSouth India. Indian J Pediatr 1994;61:223–9.

21. Walravens PA, Chase P. The prognosis of childhood systemiclupus erythematosus. Am J Dis Child 1976;130:929–33.

Received for publication 2 July 2001Accepted in revised form 14 April 2002

Mortality in Childhood Lupus 465