Effect of cotrimoxazole on causes of death, hospital admissions and antibiotic use in HIV-infected children

 Philosophie

 2 views
of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Description
Effect of cotrimoxazole on causes of death, hospital admissions and antibiotic use in HIV-infected children
Share
Tags
Transcript
  Copyright © Lippincott Williams Wilkins. Unauthorized reproduction of this article is prohibited. Effect of cotrimoxazole on causes of death, hospitaladmissions and antibiotic use in HIV-infected children Veronica Mulenga a , Deborah Ford b , A. Sarah Walker b , DarlingtonMwenya a , James Mwansa a , Frederick Sinyinza a , Kennedy Lishimpi a ,Andrew Nunn b , Stephen Gillespie c , Ali Zumla d , Chifumbe Chintu a ,Diana M. Gibb b and the CHAP Trial Team  Background:  Cotrimoxazole prophylaxis reduces morbidity and mortality in HIV-1-infected children, but mechanisms for these benefits are unclear. Methods:  CHAP was a randomized trial comparing cotrimoxazole prophylaxis withplacebo in HIV-infected children in Zambia where background bacterial resistance tocotrimoxazole is high. We compared causes of mortality and hospital admissions, andantibiotic use between randomized groups. Results:  Of 534 children (median age, 4.4 years; 32% 1–2 years), 186 died and 166had one or more hospital admissions not ending in death. Cotrimoxazole prophylaxiswas associated with lower mortality, both outside hospital ( P  ¼ 0.01) and followinghospital admission ( P  ¼ 0.005). The largest excess of hospital deaths in the placebogroup was from respiratory infections [22/56 (39%) placebo versus 10/35 (29%)cotrimoxazole]. By 2 years, the cumulative probability of dying in hospital from aserious bacterial infection (predominantly pneumonia) was 7% on cotrimoxazole and12%onplacebo( P  ¼ 0.08).Therewasatrendtowardsloweradmissionratesforseriousbacterial infections in the cotrimoxazole group (19.1 per 100 child-years at risk versus28.5 in the placebo group,  P  ¼ 0.09). Despite less total follow-up due to highermortality, more antibiotics (particularly penicillin) were prescribed in the placebogroupinyearone[6083comparedto4972daysinthecotrimoxazolegroup( P  ¼ 0.05)]. Conclusions:  Cotrimoxazole prophylaxis appears to mainly reduce death and hospitaladmissions from respiratory infections, supported further by lower rates of antibioticprescribing.AssuchinfectionsoccurathighCD4cellcountsandarecommoninAfrica,the role of continuing cotrimoxazole prophylaxis after starting antiretroviral therapyrequires investigation.    2007 Lippincott Williams & Wilkins  AIDS  2007,  21 : 77–84 Keywords: Paediatric, HIV, Africa, cotrimoxazole, prophylaxis Introduction Clinical trials and observational studies conducted inadults and children in Africa, including Cote d’Ivoire,South Africa, Malawi, Zambia and Uganda, havereported reductions of 25–46% in mortality as well asreductions in morbidity and hospital admissionsassociated with cotrimoxazole prophylaxis [1 – 6], evenin areas of high background resistance. In 2004 results of the CHAP randomized placebo-controlled trial inZambian children, showed a 43% reduction in mortalityand a 23% reduction in hospital admissions across allages and levels of CD4 cell percentage [7]. At the time,we briefly discussed the potential mechanisms of action From the  a University Teaching Hospital, Lusaka, Zambia, the  b MRC Clinical Trials Unit, London, the  c Royal Free Hospital,London, and the  d University College London, London, UK.Correspondence to Dr A Sarah Walker, Medical Research Council Clinical Trials Unit, 222 Euston Road, London NW1 2DA, UK.E-mail: asw@ctu.mrc.ac.uk * See Appendix.Received: 7 June 2006; revised: 25 August 2006; accepted: 26 September 2006. ISSN 0269-9370  Q  2007 Lippincott Williams & Wilkins 77  Copyright © Lippincott Williams Wilkins. Unauthorized reproduction of this article is prohibited. of cotrimoxazole prophylaxis, which were not entirelyclear in an African country such as Zambia with highand increasing background rates of bacterial resistanceto the drug (J. Mwansa, personal communication; for example, 84 and 75% resistance to cotrimoxazole of non-typhoidal salmonellae [8] and  Streptococcus pneumo-niae   [9]). In particular, we stressed that  Pneumocystis jiroveci   pneumonia was unlikely in this trial of childrenafter infancy as all except one of the nasopharyngealaspirates (NPAs) taken from hospitalized children withrespiratory infections were negative for   P. jiroveci   both byimmunofluorescence and polymerase chain reaction(PCR) [7].In this paper, we report causes of death, of hospitaladmissions not ending in death, and antibiotic use,amongst children randomized to cotrimoxazole prophy-laxis compared with placebo in the CHAP trial. Methods Details of the design and main results of the trial havebeen published elsewhere [7]. In brief, CHAP was arandomized placebo-controlled single centre trial inHIV-infected children in Lusaka, Zambia. BetweenMarch 2001 and January 2003, 534 HIV-infectedchildren were randomized to receive cotrimoxazoleprophylaxis or matching placebo. The trial was stoppedearly (October 2003), after a median follow-up of 19 months due to substantial and statistically significantreductions in mortality in the cotrimoxazole group.Where a child died in hospital, causes of death wereassigned by the treating paediatrician, or if the child diedbefore seeing a paediatrician, by trial clinical officers,based on information from hospital notes and nursingstaff. Subsequentlyall hospital deaths were reviewed and aprimary cause of death assigned by two paediatricians,blind to assignment to cotrimoxazole or placebo, whohad access to case report forms, death certificates, hospitalnotes, post-mortem findings and laboratory data.Agreement between reviewed and srcinally reportedcauses was relatively high with the notable exception of tuberculosis (reported as a cause for 17 children, primaryagreed cause in only four of 17).Hospitaladmissioncasereportformswerevalidatedagainsthospital notes by an independent trial manager from theMRC Clinical Trials Unit, and a trained data monitor inLusaka. Investigations were performed on childrenadmitted to hospital wherever possible. These includedNPA samples for children with suspected respiratoryinfections, blood cultures, chest X-rays, thick/thin filmsfor malaria parasites, and cerebro-spinal fluid (CSF), urineand other specimens taken for microscopy and culturewhere appropriate. Children with suspected tuberculosishad NPAs, sputa and/or gastric washings examined for acid-fast bacilli wherever possible. All NPAs wereexamined for   P. jiroveci   by immunofluorescence, andsubsequently were tested for   P. jiroveci   using PCR in theUK. Autopsies were requested from parents or carers if achild died, but only 12 were performed. Serious bacterialinfections were defined as presumptive or definitivebacterial infections of an internal organ (e.g., septicaemia,meningitis, septic arthritis, pneumonia, empyema) requi-ring hospitalization.Laboratory investigation of 13 blood cultures, 34 malariafilms and five CSF specimens were undertaken in themonth prior to death on 40 children who were admittedto hospital and subsequently died. Ninety-eight bloodcultures, 166 malaria films and five CSF specimens wereinvestigated in the month preceding or during 187hospital admissions on 132 children admitted to hospitalwho survived the admission. In total 119 NPAs wereundertaken in 107 children with respiratory symptoms,most during hospital admission. Statistical analysis Proportions were compared by Fisher’s exact tests. Daysin hospital or on antibiotics were compared using theWilcoxon rank-sum test. Cause-specific hazard ratioswere estimated by Cox proportional hazards regression.The cumulative incidences of death from different causeswere calculated from cause-specific hazards and overallsurvival curves within a competing risks framework [10].Hazard ratios based on cumulative incidence were similar to the cause-specific hazard ratios presented (data notshown). Hospital admission rates were estimated by thetotal number of admissions per child-year of follow-up,censoring follow-up at death outside hospital, hospitaladmissionifthechilddiedasaninpatient(toavoidoverlapbetween analyses of cause of death and cause of admission), loss to follow-up (8 weeks after last clinicvisitifnolongerattending)orcloseofthetrial,whichever occurred first. Statistical significance was assessed usingWald tests.CD4 cell count results were expressed as percentage of total lymphocyte counts. Height and weight wereexpressed as age-adjusted  z -scores with reference toUKstandardsfor uninfectedchildren[11].Baselinevaluesand values prior to death or hospital admission were thoserecorded nearest to randomization or the event, butbefore andwithin 6 weeks (12weeksfor CD4cell count).All analyses were performed using Stata version 8.2 (StataCorp., College Station, Texas, USA). Results At trial entry, the median age of the 534 children was 4.4 years, with 36% under 2 years (only 4% < 1 year and all 78 AIDS  2007, Vol 21 No 1  Copyright © Lippincott Williams Wilkins. Unauthorized reproduction of this article is prohibited. over 6 months), 28% 2–5 years, 21% 6–9 years and 15%over 10 years. Overall 74 (28%) children died in thecotrimoxazole group and 112 (42%) died in the placebogroup [hazard ratio (HR) 0.57; 95% confidence interval(CI), 0.43–0.77; Table 1]. As reported previously,mortality was reduced across all ages (heterogeneity P  ¼ 0.82) and across levels of baseline CD4 cellpercentage ( P  ¼ 0.36) [7]. Ninety-one children diedin hospital, 92 at home, two on the way to hospital, andone in a local clinic. Children receiving cotrimoxazolehad similar reductions in the instantaneous risk of hospital death [cause-specific hazard ratio (CHR) 0.55;95% CI, 0.36–0.84] and non-hospital death (CHR0.60; 95% CI, 0.40–0.90) compared with the placebogroup (Table 1).There were 368 hospital admissions in total (described byChintu  et al  . [7]), of which 91 were admissions prior todeath in hospital. We consider here the remaining 277admissions in 166 children; 135 admissions in 76 childrenon cotrimoxazole and 142 admissions in 90 children onplacebo. The rate of hospital admissions not ending indeath was 41 per 100 child-years at risk in thecotrimoxazole group and 53 in the placebo group[incidence rate ratio (IRR) 0.78; 95% CI, 0.55–1.10; P  ¼ 0.16; Table 1]. The number of days spent in hospitalwas higher in the placebo group (1043 days) than in thecotrimoxazole group (881 days;  P  ¼ 0.03) but individuallengths of hospital stay were similar. There was noevidence of a difference in the effect of cotrimoxazole onhospital admission rates across age at trial entry(heterogeneity  P  ¼ 0.90) or baseline CD4 cell percentage( P  ¼ 0.62) (data not shown). Deaths by cause A primary cause of death was assigned for all but four of the children who died in hospital (Table 2). Seriousbacterialinfections,predominantlypneumonia,weretheleading designated cause of death (45/91), followed bydiarrhoea (12/91) and malnutrition (7/91). However, of the 91 children who died in hospital, 43 (47%) diedonthedayofadmissionorthefollowingday:themajorityof causes (45) could therefore only be determinedpresumptively with only 15 having laboratory confir-mationofdiagnosis,andafurther27consideredclinicallydefinitive. In particular, four of the five children whodied from sepsis had blood culture results:  Haemophilusinfluenzae  (1), Salmonella spp.(1)andnogrowth(2).Onlytwo of eight children who died from clinical meningitishad a definitive diagnosis ( Streptococcus pneumoniae  cultured from CSF (1); both diagnoses confirmed atautopsy).The largest imbalance in causes of hospital death betweenrandomized groups was pneumonia or empyema (mostlydiagnosed presumptively), accounting for 10 of 35 (29%)of hospital deaths in children on cotrimoxazole compared Cotrimoxazole in African children  Mulenga  et al  .  79 Table 1. Mortality and hospital admissions not ending in death. Cotrimoxazole Placebo CHR or IRR P  ( n ¼ 265) ( n ¼ 269) (95%CI)DeathsHospital deaths 35 (13%) 56 (21%) 0.55 (0.36–0.84) 0.005Deaths outside hospital 39 (15%) 56 (21%) 0.60 (0.40–0.90) 0.012All deaths 74 (28%) 112 (42%) 0.57 (0.43–0.77)  < 0.001Hospital admissions not ending in death  1 admission 76 (29%) 90 (33%) – 0.26Admission rate per 100 child years at risk 41.0 52.5 0.78 (0.55–1.10) 0.16Total days in hospital per child (median, IQR)All children 0 (0–2) 0 (0–4) – 0.03Children admitted 7 (4–15) 8 (4–14) – 0.93Total days in hospital 881 1043 – 0.03CHR, cause-specific hazard ratio; CI, confidence interval; IQR, interquartile range; IRR, incidence rate ratio. Table 2. Primary cause of death in hospital. Deaths (with cause confirmed)Cause of death Cotrimoxazole Placebo TotalSepticaemia 3 (2) 2 (0) 5 (2) a Meningitis 4 (2) 4 (0) 8 (2) b Pneumonia/ empyema10 (3) 22 (6) 32 (9) c Serious bacterialinfections17 (7) 28 (6) 45 (13)Other infections e 1 2 3Tuberculosis 1 3 (1) 4 (1) c Diarrhoea and/ordehydration4 8 12Malnutrition f  3 4 7Malaria 1 1 (1) 2 (1) d Anaemia 2 1 3Other g 5 6 11Not known 1 3 4All hospital deaths 35 (7) 56 (8) 91(15) a Blood cultures ( Salmonella  spp.,  Haemophilus influenzae ). b Cerebro-spinal fluid culture ( Streptococcus pneumoniae ) and post-mortem (1), post-mortem (1). c Post-mortem results. d Malaria film. e Measles (2), chicken pox (1). f  Marasmus only(1), kwashiorkor only(2), marasmickwashiorkor (4). g Encephalitis (3), renal failure (2), cardiac failure (3), cor pulmonale(1), Kaposi sarcoma (1), infected burns (1).  Copyright © Lippincott Williams Wilkins. Unauthorized reproduction of this article is prohibited. with 22 of 56 (39%) of hospital deaths on placebo( P  ¼ 0.37).By2years,thecumulativeprobabilityofdyingin hospital from a serious bacterial infection was 7% in thecotrimoxazole group and 12% in the placebo group( P  ¼ 0.08, adjusted for competing risks due to other causes of death). Corresponding probabilities for hospitaldeath due to other primary causes were 8 and 13%( P  ¼ 0.12); and 18 and 26% for death outside hospital( P  ¼ 0.04). There was no evidence that mortalityreductions in the cotrimoxazole group varied by cause(serious bacterialinfections versusother) or place of death(heterogeneity  P  ¼ 0.96).For the 95 children who died outside of hospital, cause of death is uncertain, and was ascertained through verbalreports from relatives by a team nurse or clinical officer who visited the home. The majority were reported tohave had diarrhoea and/or dehydration (30) or respir-atory problems (20); for 27 children no informationwas available. Hospital admissions by cause Diagnosesathospitaladmissionfollowedasimilarpatternto hospital deaths (Table 3) with pneumonia/empyemabeing the most common (127 admissions). Multiplediagnoses were common, with at least two distinctdiagnoses reported for 122 admissions (44%). Presumedbacterial pneumonia/empyema was frequently diag-nosed with other conditions, being a concurrentdiagnosis in 47, 31 and 24% of admissions for tuberculosis, malaria and malnutrition. As for cause of death, definitive diagnosis with laboratory confirmationwas infrequent (Table 3).  Staphylococcus aureus  and Salmonella  species were the main isolates from bloodculture; neither   S. pneumoniae   nor   H. influenzae   wasisolated from blood cultures although prevalence of carriage from pernasal swabs taken from children in aplanned substudy throughout the trial was 51 and 29%respectively [9]. Only four of the 55 diagnoses of malariaat admission were confirmed by the presence of parasitaemia; in the remainder, diagnosis was based on 80 AIDS  2007, Vol 21 No 1 Table 3. Hospital admissions not ending in death. Admissions (number of children)[diagnosis confirmed]Admission rate per 100child-years at riskAdmissions with at leastone other diagnosis a Diagnosis Cotrimoxazole Placebo Total Cotrimoxazole Placebo IRR (95% CI)  P   (%)Septicaemia 7 (7) 7 (7) 14 (14) [11] b 2.1 2.6 0.82 (0.29–2.32) 0.71 14 (100%)Meningitis 3 (3) 2 (2) 5 (5) 0.9 0.7 1.23 (0.21–7.36) 0.82 1 (20%)Pneumonia/ empyema56 (39) 71 (47) 127 (86) [2] c 17.0 26.3 0.65 (0.40–1.04) 0.07 67 (53%)Serious bacterialinfections f  63 (42) 77 (52) 140 (94) 19.1 28.5 0.67 (0.42–1.06) 0.09 76 (54%)Respiratory(non-serious)infections22 (19) 20 (16) 42 (35) (34) 6.7 7.4 0.90 (0.45–1.82) 0.78 26 (62%)Tuberculosis 20 (18) 27 (25) 47 (43) [4] d 6.1 10.0 0.61 (0.33–1.12) 0.11 41 (87%)Malaria 32 (25) 23 (17) 55 (42) [4] e 9.7 8.5 1.14 (0.59–2.21) 0.70 41 (75%)Diarrhoea and/ordehydration26 (17) 17 (15) 43 (32) 7.9 6.3 1.25 (0.57–2.75) 0.57 25 (58%)Other infections g 8 (7) 9 (9) 17 (16) 2.4 3.3 0.73 (0.27–1.99) 0.54 11 (65%)Neurological h 2 (2) 4 (4) 6 (6) 0.6 1.5 0.41 (0.08–2.22) 0.30 3 (50%)Malnutrition i 14 (13) 19 (15) 33 (28) 4.2 7.0 0.60 (0.27–1.35) 0.22 24 (73%)Anaemia 4 (2) 5 (5) 9 (7) 1.2 1.8 0.66 (0.11–3.88) 0.64 9 (100%)Other  j 20 (13) 13 (12) 33 (25) 6.1 4.8 1.26 (0.54–2.97) 0.60 18 (55%)Not known 1 (1) 3 (3) 4 (4)Total 135 (76) 142 (90) 277 (166) [21] 41.0 52.5 0.78 (0.55–1.10) 0.16 122 (44%)Child years at risk 330 270CI, confidence ratio; IRR, incidence rate ratio. a Distinctdiagnosesaslistedinthistablewiththeexceptionthat‘Other’(  j )wassplitoutintoseparatediagnoses;thatisachildadmittedfortwonon-respiratory infections would not be considered as a multiple diagnosis. b Blood cultures: cotrimoxazole group (6): 3  Salmonella  spp., 2  Staphylococcus aureus , 1  Klebsiella pneumoniae , placebo group (5): 1  Salmonella spp., 3  S .  aureus , 1  Proteus mirabilis . c Blood culture: cotrimoxazole group (1):  Klebsiella pneumoniae ; NPA: placebo group (1):  Pneumoniae jiroveci  . d Gastric washings/sputum: cotrimoxazole group (1), placebo group (3). e Malaria film: cotrimoxazole group (1), placebo group (3). f  Subtotal for admissions (children) with serious bacterial infections will not be the sum of admissions for septicaemia, meningitis and pneumonia/ empyema because there were some admissions with  > 1 diagnosis. g Measles (1), chicken pox (7), skin (5), urinary (2), conjunctivitis (2). h Encephalitis (1), peripheral neuropathy (2), spasticity (3). i Marasmus only (15), kwashiorkor only (4), marasmic kwashiorkor (12), unspecified (2).  j Cardiacfailure(4),arthritis/arthralgia(3),skin(2),rectalbleeding(1),epistaxis(1),auricularcyst(2),parotitis(2),oralthrush(10),fevers(1),Kaposisarcoma (9), pleural effusion (1); three admissions with two diagnoses.  Copyright © Lippincott Williams Wilkins. Unauthorized reproduction of this article is prohibited. history and response to antimalarial drugs. Four of 47cases of tuberculosis were confirmed by gastric washings(1) or sputum (3). Of 119 NPAs from 107 children withrespiratory symptoms, only one had  P. jiroveci   identifiedon immunofluorescence. The child, aged 9, was onplacebo and responded to high-dose cotrimoxazoletherapy. Following recovery she received open labelcotrimoxazole prophylaxis. Retrospective testing of 73NPA samples for   Pneumocystis  by PCR were all negative.The serious bacterial infections admission rate was 19.1per 100 child-years at risk in the cotrimoxazole groupand28.5intheplacebogroup(IRR0.67;95%CI,0.42– 1.06;  P  ¼ 0.09). A similar reduction in the admission ratefor malnutrition was observed (IRR 0.60; 95% CI,0.27–1.35;  P  ¼ 0.22), but this was based on smaller numbers and did not reach statistical significance.Although there was an apparent reduction in theadmission rate for tuberculosis (IRR 0.61; 95% CI,0.33–1.12;  P  ¼ 0.11), admission rates in the treatmentgroups were very similar restricting to the 32 admissionswith anti-tuberculosis treatment recorded at the time(IRR 0.93; 95% CI, 0.44–1.95;  P  ¼ 0.85); in six of 15admissionswithnotreatmentrecorded,pneumoniawasaconcurrent diagnosis.By 2 years, the cumulative probabilityof first admission toor death in hospital with serious bacterial infections was15% in the cotrimoxazole group and 25% in the placebogroup ( P  ¼ 0.05) (Fig. 1). Comparative probabilities for other causes of first admission to or death in hospital were24 and 28% ( P  ¼ 0.27); and 12 and 17% for death outsideof hospital ( P  ¼ 0.15). The risk of first admission or hospital death with serious bacterial infections was 41%lower in the cotrimoxazole group, and without seriousbacterial infections 28% lower (CHR ¼ 0.59; 95% CI,0.39–0.89 and CHR ¼ 0.72; 95% CI, 0.50–1.04,respectively, heterogeneity  P  ¼ 0.49). Age, CD4 cell percentage, weight and height attime of death or hospital admission Underlying chronic malnutrition was prevalent in thisgroup of children [7]; at trial entry height and weight-for-age were below the third centile in 78 and 67%children respectively; body mass index-for-age was lessthan the third centile in 22%. Although numbers weresmall there was a suggestion that children with lowestweight-for-age contributed more hospital deaths andhospital admissions in the placebo group. This was largelydriven by pneumonia/empyema, for which the pro-portion of deaths, first and all hospital admissions inchildren with weight-for-age scores of –4 or less was 81,48 and 40% in the placebo group compared with 25, 27and 30% in the cotrimoxazole group ( P  ¼ 0.02, 0.16 and0.31, respectively). There were no differences in thedistribution of age, height-for-age or CD4 cell percen-tage among deaths or hospital admissions betweenchildren on cotrimoxazole and placebo. Use of antibiotics and anti-tuberculosis therapy Most children (84%) were treated with short-termantibiotics for infections at some time during the trial.Total dayson antibioticswere higher intheplacebogroup(7952 days) than in the cotrimoxazole group (7401 days, P  ¼ 0.68) despite a shorter period of follow-up in theplacebo group due to poorer survival. As expected, under informative censoring whereby the sickest children die Cotrimoxazole in African children  Mulenga  et al  .  81 051015202530    C  u  m  u   l  a   t   i  v  e   i  n  c   i   d  e  n  c  e   (   %   ) 2.01.51.00.50.0    C  u  m  u   l  a   t   i  v  e   i  n  c   i   d  e  n  c  e   (   %   ) 1.52.01.00.50.0 Years since randomisation    C  u  m  u   l  a   t   i  v  e   i  n  c   i   d  e  n  c  e   (   %   ) 0510152025300.00.51.0 1.5 2.0 Non hospital death, activeNon hospital death, placebo 051015202530 (a)(b)(c) Other HA/death, activeOther HA/death, placeboSBI HA/death, activeSBI HA/death, placebo Fig.1. Cumulativeincidenceoffirsthospitaladmission(HA)or death in the presence of other risks in active and placebogroups, from top: (a) hospital admission or hospital death,serious bacterial infection (SBI); (b) hospital admission orhospital death, other cause; (c) death outside hospital.
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks