INICIATIVA DEL PODER EJECUTIVO DE

From the prepared RNA, conventional reverse-transcription polymerase chain reaction was carried out. A 348-bp fragment of the S-RNA was amplified by use of the Platinum Thermoscript One-step System (Life Technologies) in a 20-µL assay containing 2 µL of RNA, 0.3 µM primer 36E2, and 0.2 µM primer 80F2 or alternatively, 0.2 µM primer 80F or alternatively 0.2 µM primer 80F2-CSF (ATGTAGTGGTGACTAGTATTTTTTGTGCA;

modification to 80F2 underlined).

The modification was run in a thermocycler (model 9600; Perkin Elmer) as follows: 50 ⁰C for 30min; 95 ⁰C for 5 min; 10 step-down cycles at 95 ⁰C for 5s, 60 ⁰C for 5s with a 1⁰C decrease per cycle, and 72 ⁰C for 25s, 40 cycles at 95 ⁰C for 5s, 56 ⁰C for 10s, 72 ⁰C for 25s and a final 5 min at

72⁰C. The ≥ 95% detection limit of the PCR was 2.5 x 10³ RNA copies/mL. ^{63, 110}

Packed cell volume (PCV) and White Blood Cell count (WBC).

Capillary centrifugation method was used to measure PCV using a Hawksley microhaematocrit centrifuge and reader. Dilution fluid was 0.I N Hydrochloric acid. Blood films were prepared using a method previously described by Dacie and Lewis. ¹¹⁹ The WBC count was afterwards performed using the improved Neubauer counting chamber.

Blood film for Malaria Parasites

53

Thick and thin blood films for plasmodium count and specie identification were prepared respectively. The Giemsa technique ¹¹⁹ was employed for staining blood films. The author read the slides which were cross checked by a Medical Laboratory Scientist at the Department of Medical Microbiology, ISTH. Parasite numbers per l were estimated by counting the number of asexual parasites against the number of leucocytes in the thick blood film. ¹¹⁹

Blood culture

Blood culture was done as previously described. ⁷ Blood culture bottles were incubated for seven (7) days before being declared sterile. The sensitivity patterns of isolates were determined through the aid of multidiscs impregnated with antibiotics. Bacteraemia was defined as the presence of bacterial pathogen in the blood on culture. ⁸

Random plasma glucose

Sodium fluoride oxalate sample bottles were used for collecting venous blood and the glucose level estimated by the enzymatic glucose oxidase method.¹²⁰ A random plasma glucose of less than 2.2 mmol/L or 40 mg/dl was considered indicative of hypoglycaemia.¹²¹

Lumbar puncture (LP)

LP was done through the intervertebral space between the third and fourth lumbar vertebrae. Patients were placed in the left lateral position and the skin prepared aseptically using chlorhexidine soaked sterile gauze to cleanse the area in concentric circles enlarging from the inside outwards initially followed by cleansing with 70% alcohol.

The CSF samples thus obtained were used for biochemical and microbiological analysis.^116,

121

A diagnosis of pyogenic meningitis was made on the basis of CSF pleocytosis ( CSF WBC count greater than or equal to 10 cells / mm³, greater than 60% polymorphs ), in addition to the identification of bacteria by Gram stain and/or CSF or blood culture. ⁵⁶

54

In the absence of positive bacteria identified either by Gram stain results or negative CSF or blood culture, especially in those pretreated with antibiotics prior to presentation, the diagnosis of pyogenic meningitis was made in the presence of CSF pleocytosis ( WBC count greater than or equal to 10/mm³, greater than 60% polymorphs ) and typical biochemical changes namely, CSF protein greater than 80mg % and CSF glucose less than 50% of concomitant random plasma glucose. These patients were diagnosed as having probable meningitis. ^{40, 42}

Urinalysis and Microscopic urinalysis.

Urine was collected in a universal sterile bottle by clean catch method (in older toilet-trained children) and by suprapubic aspiration in infants and very ill children. One drop of uncentrifuged urine, was delivered via a micropipette, placed on the improved Neubauer counting chamber and was subsequently examined under the microscope for bacteria and WBC

at 100x objective. ¹²¹

MANAGEMENT OF THE PATIENTS

Seizures were controlled using intravenous diazepam at a dose of 0.3 mg per kilogram body weight given slowly by titration as in a previous study.¹²² Children with multiple seizures received phenobarbitone at a loading dose of 15-20mg/kg/body weight given intramuscularly statim, and were subsequently maintained on 5 mg/kg/body weight per day in 2 or 3 divided doses, also as recommended in an earlier study. ¹²²

Children with suspected or documented hypoglycemia were given 10% dextrose in water at a recommended dose of 2-5ml/kg body weight given bolus intravenously and then maintained on intravenous 10% dextrose in water given at 6-8mg/kg/min until they were able to commence

oral feeding. ¹²³

55

Patients with LF were given intravenous Ribavirin at 33mg/kg body weight statim and subsequently maintained on 16mg/kg body weight/dose administered 6 hourly for the first 4 days. Thereafter, 8mg/kg body weight/dose of Ribavirin was administered 8 hourly for another

6 days. In all, therapy with Ribavirin was for a recommended duration of 10 days. ¹²⁴

Children with severe malaria received intravenous quinine at a recommended loading dose of 20 mg per kilogram body weight given in 10ml per kg body weight of 5% dextrose in water as an infusion over 4 hours statim and subsequently maintained on intravenous quinine 10mg per kg per dose in 10ml per kg of 10% dextrose in water given 8 hourly, slowly over 4 hours until the patient regained consciousness fully. Intravenous quinine was at this time changed to oral, at the same dose, until 7 days of therapy was completed.

4 Those with uncomplicated malaria were treated with fixed drug combined tablets of Artemether-Lumefantrine with dosing based on body weight as follows; Children between 5-14kg received one tablet, 12 hourly for 3 days; those between 15-24kg received two tablets, 12 hourly for 3 days; 25-34kg received three tablets, 12 hourly for 3 days while those > 35kg received four tablets, 12 hourly for 3 days.

Therapy for suspected bacterial meningitis was intravenous Ceftriaxone at a dose of 100mg/kg body weight statim followed by maintenance at a dose of 100 mg/kg body weight/day once daily.¹²⁵ Ceftriaxone had been recommended as the drug of first choice in bacterial meningitis.¹²⁵ The Ceftriaxone was administered intravenously once in every 24hours over a period of 10-14 days.¹²⁵

Children suspected of having bacteraemia were given an initial antibiotic combination of intravenous Ceftriaxone at 100mg/kg body weight/day once daily and intramuscular Gentamicin at 5 mg /kg body weight in 3 equal divided doses. Antibiotic therapy was reviewed

when clinically indicated or in line with the blood culture sensitivity report.⁷

The observed outcome of illness was recorded, namely whether the patient completed treatment, discharged against medical advice or died.

56 STATISTICAL ANALYSIS

Data was stored in Microsoft Excel software and processed using Microsoft Windows Vista operating system and the significance of test results analyzed using Epi info version 3.5.1. Frequencies were compared using

χ

²-test with Yates correction or Fisher’s exact test as appropriate and means with Student’s t-test. ^{126, 127} P values less than 0.05 were regarded as significant for all the statistical analysis.

ETHICAL CONSIDERATIONS

Approval for the study was obtained from the Research and Ethics Committee of ISTH.

The study was undertaken in accord with internationally accepted standards of research involving human subjects ¹²⁹. All subjects were treated fairly following the principles of justice and equity, respect for persons, beneficience and non-malficience.

RESULTS

Nine hundred and thirteen children ( 913 ) were admitted into the CHER from 1^ST of December

57

2009 to 30^th of November 2010. Three hundred and seventy three ( 40.9% ) of these had fever. One hundred and eight ( 29.0% ) of the 373 children with fever (or 11.8% of the 913 admissions) had convulsion associated with fever (Cases), while 265 ( 71.0% ) were without convulsions (Controls).

PREVALENCE OF LASSA FEVER

Thirteen children ( 3.5% [ 95% CI = 1.9%, 5.7% ] ) of the 373 children with fever or 1.4% ( 95%

CI = 0.8%, 3.4% ) of the 913 total admissions had LF. The monthly distribution of Cases and Controls is shown in Table I. The number of Cases was lowest in April and highest in July, while the number of Controls was lowest in April and July but highest in November.

The difference between the distribution of Cases and Controls was highly significant (

χ

² = 36.27, df = 3, p <0.0001).

The monthly distribution of cases of LF in relation to the monthly distribution of children with fever is shown in Table II and Figure 1. The prevalence of LF amongst febrile children was highest in June. However, there was no significant difference between the prevalence of LF in febrile children in November to April (dry season) (6/178 or 3.4%) and May to October (rainy season) (7/195 or 3.6%) (

χ

² = 0.028; Odd’s ratio (OR) [95% CI] = 0.9 [ 0.31, 2.84 ]; p = 0.867).

Table I: Monthly distribution of Cases and Controls.

Month

Cases (n = 108 ) No. (%)

Control (n = 265 ) No. (%)

Total (n = 373) No. (%)

No. (%) with LF

58 December

January February March April May June July August September October November

8 (7.4) 6 (5.6) 5 (4.6) 6 (5.6) 1 (0.9) 3 (2.8) 13 (12.0) 32 (29.6) 9 (8.3)

7 (6.5) 4 (3.7) 14 (13.0)

41 (15.5) 12 (4.5) 19 (7.2) 12 (4.5) 5 (1.9) 6 (2.2) 21 (8.0) 5 (1.9) 26 (9.8) 23 (8.7) 46 (17.3) 49 (18.5)

49 (13.1 ) 18 (4.8 ) 24 (6.4 ) 18 (4.8 )

6 (1.6 ) 9 (2.4 ) 34 (9.1 ) 37 (10.0 ) 35 (9.4 ) 30 (8.0 ) 50 (13.4 ) 63 (17.0 )

3 (23.1 ) 0

0

2 (15.4 ) 0

0

4 (30.8 ) 2 (15.4 ) 0

0 1 (7.7 ) 1 (7.7 ) Total 108 (100.0) 265 (100.0) 373 (100.0) 13 (100.0)

χ

² = 36.27; df = 3 (January to March vs April to June vs July to September vs October to December); p < 0.0001 for distribution of Cases vs Controls .

Table II: Monthly prevalence of Lassa fever in febrile children.

No. of febrile children No. with LF ( % )

59 Month

December January February March April May June July August September October November

49 (13.1 ) 18 (4.8 ) 24 (6.4 ) 18 (4.8 )

6 (1.6 ) 9 (2.4 ) 34 (9.1 ) 37 (9.9 ) 35 (9.4 ) 30 (8.0 ) 50 (13.4 ) 63 (16.9 )

3 (6.1 )

0 0 2 (11.1 ) 0 0 4 (11.8 ) 2 (5.4 ) 0 0 1 (2.0 ) 1 (1.6 ) Total 373 (100.0) 13 (3.5)

χ

² = 36.27, df = 2 , p = 0.264

60

0 % 10 % 20 % 30 %

% 40

% 50 60 % 70 % 80 % 90 % 100 %

Dec to Feb

Mar to May

Jun to Aug

Sep to Nov

LF

Febrile children

61

Figure 1: Quarterly prevalence of Lassa Fever in febrile children.

PREVALENCE OF LASSA FEVER AND OTHER INFECTIONS IN CASES AND CONTROLS

The contributions of LF, malaria parasitaemia, extra-cranial focal infections and meningitis to the diagnoses in Cases and Controls is shown in Table III. Overall, 13 (3.5 % [95% CI = 1.9%, 5.7%]) of the febrile children had LF. The prevalence of LF was 3/108 (2.8%; [95% CI

= 0.7%, 7.4 %]) among the Cases and 10/265 (3.8%; [95% CI = 1.9%, 6.6%]) among the controls. The difference was not statistically significant (OR *95% CI+ = 0.73 *0.2, 2.7+, Fisher’s exact p =

0.903).

The Cases had a significantly higher prevalence of malaria parasitaemia (OR [95% CI] = 4.90 [2.92, 8.23], p 0.0001) and a significantly lower prevalence of focal extra-cranial infections (FECI) (OR [95% CI] = 0.47 [0.24, 0.91] and undetermined infections (OR [95% CI] = 0.40 [0.25, 0.66]. Plasmodium falciparum was the specie in all 87 children with malaria parasitaemia. The higher prevalence of meningitis among the Cases approached significance (OR [95% CI] = 2.60 [1.05, 6.44], p = 0.060). Differences in the prevalence of miscellaneous diagnoses were not

significant ( p > 0.05 ).

Four (30.8%) of the 13 children with LF had co-infection with malaria parasitaemia. Malaria coinfection was present in 3/3 cases and 1/10 controls with LF (Fisher’s exact p = 0.028).

Table III: Pattern of diagnoses in Cases and Controls.

62

Cases (n = 108) No. (%)

Controls (n = 265) No. (%)

χ

^{2 / p} Odds ratio ( 95%

C I )

Diagnosis

Lassa fever with malaria ^* Lassa fever only

Malaria only ^*

Focal extra-cranial infections Meningitis

Miscellaneous Undetermined

3 ( 2.8 ) 0 47 ( 43.5 ) 12 ( 11.1 ) 10 ( 9.3 )

9 ( 8.3 ) ^α 27 ( 25.0 )

1 ( 0.4 ) 9 ( 3.4 ) 36 ( 13.6 )

56 ( 21.1 ) 10 ( 3.8 ) 33 ( 12.5 ) ^β 120 ( 45.3 )

- / 0.150^**

-/0.089^**

38.03 / 0.0001 4.52 / 0.034 3.53 / 0.060 0.92 / 0.337^**

12.38 / < 0.001

7.54 (0.78, 73.33)

4.90 (2.92, 8.23) 0.47 (0.24, 0.91) 2.60 (1.05, 6.44) 0.64 (0.30, 1.39) 0.49 (0.25, 0.66)

Total

108 (100)

265 (100)

χ

² = 48.81 ; df = 5 (Lassa fever with malaria merged with Lassa fever only) ; p 0.0001

* Malaria parasitaemia, ** Fisher’s exact test, α Diabetic Ketoacidosis 1, Gamalin 20 poisoning 1, HIV / AIDS 3, lymphoma 1, Myocarditis 2, Pertusis 1, β Cholestatic Jaundice 1, Chronic Renal Failure 1, Gamalin 20 poisoning 1, Gastroenteritis 15, HIV / AIDS 4, HIV/ AIDS – TB co-infection 3, Kawasaki disease 1, Marasmus 2, Metastatic Neuroblastoma 1, Pelvic Inflammatory

Disease 1, URTI 3

63

The prevalence of LF in Cases and Controls is further shown in Table IV and Figures 2 and 3, in terms of the diagnoses in relation to the type of CAWF (febrile convulsions versus CAWF excluding febrile convulsions). Irrespective of the type of CAWF, there was a significant difference between the Cases and Controls ( p = 0.0001 for

χ

²with 4 degrees of freedom for febrile convulsions versus Control group I [ Controls aged > 6 month to 5 yrs old ]and p < 0.0001 for CAWF excluding febrile convulsions versus Control group II [ Controls aged 6 month or 5yrs old ] ) in the pattern of diagnoses.

The prevalence of LF was not significantly different between Cases with febrile convulsions and those with CAWF excluding febrile convulsions ( 1/61 versus 2/47; OR [95% CI] = 0.38 [0.03, 4.27] , Fisher’s exact p= 0.805 ) or between the two groups of Controls ( 2/117 versus 8/148; OR [95% CI] = 0.30 [0.08, 1.44], Fisher’s exact p = 0.200 ). Also, there was no significant difference between the prevalence of LF in febrile convulsions versus Control group I ( 1/61 versus 2/47; OR [95% CI] = 0.96 [0.09, 10.78] ) or between Cases with CAWF excluding febrile convulsions and Control group II ( 2/47 versus 8/148; OR [95% CI] = 0.78 [0.16, 3.80] )

Compared with Control group I, the Cases with febrile convulsions had a significantly higher prevalence of malaria parasitaemia ( 31/61 versus 21/117; OR [95% CI] = 4.70 [2.37, 9.41];

χ

^{2 =}

19.39, p = 0.0001 ) and a lower prevalence of focal extra-cranial infections (FECI) ( 8/61 versus 42/117; OR [95% CI] = 0.27 [0.12, 0.62];

χ

² = 9.21, p = 0.002 ). There was no significant difference in the prevalence of miscellaneous diagnoses ( 4/61 versus 13/117;

OR [95% CI] =

0.56 [0.18, 1.80] ) and undetermined diagnoses ( 17/61 versus 39/117; OR [95% CI] = 0.77 [0.39, 1.52] ). Compared with Control group II, the Cases with CAWF excluding febrile convulsions had a significantly higher prevalence of malaria ( 16/47 versus 15/148; OR [95% CI] = 4.58 [2.95,

10.24];

χ

² = 13.51, p = 0.0002 ) and meningitis ( 10/47 versus 10/198; OR [95% CI] = 3.73 [1.44, 9.63+; Fisher’s exact p = 0.014 ) and a lower prevalence of undetermined diagnoses ( 10/47 versus 81/148; OR [95% CI] = 0.22 [0.10, 0.48];

χ

² = 14.72, p = 0.0001 ). There was

64

no significant difference in the prevalence of FECI ( 4/47 versus 4/148; OR [95% CI] = 0.89 [0.28, 2.85] ) and miscellaneous diagnoses ( 3/47 versus 20/148; OR [95% CI] = 0.76 [0.27, 2.16] ).

65

malaria only

LF with malaria

Focal extra-cranial infection miscellaneous

undetermined

66

Figure 2a: Diagnoses in Cases with febrile convulsions.

Figure 2b: Diagnoses inControls aged > 6 month to 5 yrs old ( control group I ).

malaria only

LF only

Focal extra-cranial infection

miscellaneous

undetermined

67

Figure 3a: Diagnoses in all children with CAWF excluding febrile convulsions.

malaria only

LF with malaria

Focal extra-cranial infection meningits

suspected meningitis

miscellaneous

undetermined

malaria only

LF with malaria

LF only

Focal extra-cranial infection meningits

miscellaneous

undetermined

68

Figure 3b: Diagnoses in Controls aged 6 months or 5yrs old ( Control group II ).

CHARACTERISTICS OF CONVULSIONS IN CHILDREN WITH LASSA FEVER

Overall, 3/13 (23.1%; [95% CI = 6.2%, 50.9%]) children with LF and 105/360 (29.2%; [95% CI = 24.6%, 34.0%]) with other infections/diagnoses presented with CAWF (OR [95% CI] = 0.73 [0.20, 2.7+; Fisher’s exact p = 0.903). Among the 108 Cases with CAWF, 64 (59.3%) had generalised, 13 (12.0%) focal, 11 (10.2%) tonic, 1 (0.9%) clonic and 19 (17.6%) unclassified convulsions.

Four (3.7%) children (3 with cerebral malaria and 1 with gamalin 20 poisoning) had status epilepticus. Two of the three cases with LF had generalised tonic-clonic and 1 focal convulsions.

The pattern of the convulsions in relation to diagnosis of the cause of fever is shown in Table V. There was a significant relationship between the pattern of convulsions and diagnosis (p = 0.014 for χ² with 2 degrees of freedom) . The prevalence of focal convulsions in LF (1/3) was higher than in cases with other diagnosis ( 12/105 ) but the difference was not statistically significant (OR [ 95% CI ] = 3.88 [ 0.33, 46.22 +, Fisher’s exact p = 0.644).

69

Table V: Pattern of convulsions in relation to diagnosis.

Diagnoses

Lassa Fever (n = 3) No. (%)

Malaria only (n

= 47) No. (%)

Meningitis (n = 10) No. (%)

Focal Extra Cranial Infections (n = 12) No. (%)

Miscellaneous (n = 9) No. (%)

Unidentified (n = 27 ) No. (%)

Total (n = 108 ) No. (%)

Seizure type

Generalised 2 (66.7) 33 (70.2) 9 (90.0) 6 (50.0) 5 (55.6) 9 (33.3) 64 (59.3 )

Focal 1 (33.3) 6 (12.7) 1 (10.0) 2 (8.3) 1 (11.1) 3 (14.8) 13 (12.0 )

others 0 3 (6.4) 0 2 (16.7) 1 (11.1) 6 (22.2) 12 (11.1 )

Unclassified 0 5 (10.6) 0 2 (16.7) 2 (22.2) 9 (33.3) 19 (17.6 )

Total 3 (100) 47 (100) 10 (100) 12 (100) 9 (100) 27 (100) 108 (100)

χ² (generalised convulsions versus other types of convulsion [focal seizures merged with others and unclassified seizures])

= 10.68; degrees of freedom = 3 ( LF merged with malaria, meningitis merged with focal extra-cranial infections, miscellaneous merged with unidentified ); p = 0.014

70

CLINICAL FEATURES ASSOCIATED WITH LASSA FEVER AT PRESENTATION

The prevalence of LF in relation to the age and sex distribution of the Cases and Controls is shown in Table VI. Overall, 86 / 108 (79.6% ) of Cases and 200 / 265 (75.5%) of Controls were aged

5 yrs. There was no significant difference between the Cases and Controls in their distribution with age ( p = 0.468). Among the Cases, the prevalence of LF in children 5 yrs old was not significantly higher than in older children ( 3 / 86 (3.5% ) versus 0 / 22 ; Fishers exact p > 0.999 ). Among the Controls, the prevalence of LF was also not significantly different between children 5 yrs old ( 7 / 200 or 3.5 % ) and older children ( 3 / 65 or 4.6 % ; OR [ 95%

CI ] = 0.75 [ 0.19, 2.99 ], Fisher exact p = 0.923 ). The youngest patient with LF was 6 months old while the oldest was 84 months old.

The M: F ratio was 1.4: 1 for both Cases and Controls. There was no significant difference in the prevalence of LF between males ( 9/ 215 or 4.18% ) and females ( 4 / 158 or 2.5 % ; OR [ 95% CI ] = 1.68 [ 0.51, 5.56 ], Fisher’s exact p = 0.78 ).

The prevalence of presenting symptoms and signs in Cases and Controls are shown in Tables VII and VIII. Vomiting was the most frequent symptom in both Cases ( 43.5% ) and Controls(

47.9%)

( Table VII ) and an acutely ill appearance the most frequent sign ( 87.0% in Cases and 72.1% in Controls) (Table VIII ).

Table VI: Age and sex distribution of Cases and Controls.

Cases Controls Total

(n=108 ) (n=265) (n=373)

No. (%) No. (%) No. (%)

71

No. (%) with LF No. (%) with LF No. (%) with LF

Age in years:^*

< 1 yr

≥ 1 - 5 yrs

≥ 6 - 12 yrs

> 12 yrs

Total Sex:^**

Male Female

Total

22 (20.4) 64 (59.3) 18 (16.7) 4 (3.7)

108 (100)

62 (57.4 ) 46 (42.6)

108 (100)

1 (4.6) 2 (3.1) 0 0 3 (2.8)

2 (3.2) 1 (2.2)

3 (2.8)

50 (18.9) 150 (56.6) 54 (20.4) 11 (4.1)

265 (100)

153 (57.7) 112 (42.3)

265 (100)

1 (2.0) 6 (4.0) 3 (5.6) 0

10 (3.8)

7 (4.6) 3 (2.7)

10 (3.8)

72 (19.3) 214 (57.4)

72 (19.3) 15 (4.0)

373 (100)

215 (57.6) 158 (42.4)

373 (100)

2 (2.8 ) 8 (3.7) 3 (4.2) 0 13 (3.5)

9 (4.2) 4 (2.5)

13 (3.5)

*

2

χ = 0.01 ; 2 degrees of freedom ( 6-12 yrs merged with 12yrs ) ; p = 0.994 for Cases versus Controls

** 2 χ = 0.19 ; p = 0.667 for Cases versus Controls The Cases had a significantly lower prevalence of abdominal pain ( OR [ 95% CI ] = 0.23 ( 0.09, 0.58 ), p = 0.002 ) and cough ( χ² = 7.72; OR [ 95% CI ] = 0.45 ( 0.27, 0.78 ), p = 0.005 ) ( table VII) and a significantly higher prevalence of an acutely ill appearance ( OR [ 95% CI ] = 2.60 [ 1.70, 4.85 ], p = 0.003 ) ( Table VIII ) compared to the Controls . The higher

72 prevalence of unarousable coma in Cases

approached significance (OR * 95% CI + = 2.92 * 1.10, 7.78 +, Fisher’s exact p = 0.058) ( Table VIII ). The other differences in presenting signs and symptoms between the Cases and Controls were not statistically significant.

Cases and Controls were merged in analysis for the purpose of comparison of the clinical features in children with LF with that of those without LF to make up for the small number of children with LF. The relationship between the prevalence of LF and the presence or absence of stated presenting symptoms and signs among febrile children is shown in Table IX and Figure 4 as well as Table X and Figures 5. Vomiting was associated with a lower prevalence of LF ( 1 / 174 children with vomiting versus 12 / 199 without vomiting had LF ) ; OR [ 95% CI ] = 0.09 ( 0.01,

0.70 ), p = 0.01 ) ( Table IX ).

The higher prevalence of LF in those with an acutely ill appearance did not attain statistical significance ( 13 / 285 with the sign versus 0 / 88 without the sign had LF ; Fisher’s exact p = 0.057 ) ( Table X ). Febrile children with bleeding had a statistically significantly higher prevalence of LF ( 4 / 8 with bleeding versus 8 / 365 without bleeding had LF ; OR [95% CI ] 39.56 * 8.52, 183.7 +, Fisher’s exact p = 0.0001 ). The prevalence of LF had no statistically significant relationship with any of the other symptoms or signs including dyspnoea and diarrhoea ( Tables IX and X ).

The relationship between the prevalence of LF and clinical presentation is further shown in Table XI, with respect to duration of fever, presence or absence of localising signs of infection, initial diagnosis and level of consciousness. The prevalence of LF was higher in children without localizing signs of infection although the difference did not attain statistical significance ( 13 / 289 with versus 0 / 84 without signs of localizing infection had LF ; Fisher’s exact p = 0.068 ). The prevalence of LF was higher in children in whom it was the initial diagnosis ( 9 / 11 with versus 4 / 362 without an initial diagnosis of LF had LF ) ; OR [ 95% CI ] = 402.8 ( 65.15, 2490 ), Fisher’s exact p <0.0001). There was no significant relationship between the duration of fever and presentation with unarousable coma and the prevalence of LF ( Table XI ).

73

Ninety four children ( 25.2 % ) had a history of treatment with anti-malarials and / or antibiotics before presentation. The prevalence of LF in children with a history of pre-treatment ( 5 / 94 or 5.3% ) was not significantly different from that in those without pre-treatment ( 8 / 279 or 2.9% ; OR [95% CI] = 1.9 [ 0.61, 5.97 ], p = 0.415 )

Pre-hospital treatment with cow urine concoction (CUC) and burning of the skin was not observed in any of the patients, whether Cases or Controls. However, ingestion of herbal concoctions and topical application of palm kernel oil to the skin was observed in 13 / 373 ( 3.5 % ) of the patients. Treatment with traditional remedy was not associated with hypoglycaemia in any of the patients.

74

75

76

77

Figure 4: Comparing the frequency of presenting symptom in Lassa fever, Cases vs Control.

78

79

Figure 5: The Prevalence of clinical signs at presentation in Lassa fever, Cases vs Controls.

80

81

BLOOD GLUCOSE LEVELS, HAEMATOCRIT AND WHITE BLOOD CELL COUNT IN SUBJECTS IN THE STUDY

The mean ( SD and 95% CI of the mean ) and range of random plasma glucose, haematocrit and WBC of the Cases and Controls are shown in the Table XII. There were no significant differences between the Cases and Controls in any of the three parameters.

The prevalence of abnormal blood glucose levels, severe anaemia and abnormal WBC count is shown in Table XII. Overall, 10 (2.7%) children were hypoglycaemic ( RPG < 2.2 mmol/l ) and 19 ( 5.1% ) hyperglycaemia ( RPG > 6.7 mmol/l ). The prevalence of hypoglycaemia in children with CAWF was 1.9%. None of the two children with CAWF who had hypoglycaemia received any pre-treatment with traditional remedy.

Fifty one ( 13.7% ) were severely anaemic ( haematocrit < 15% [ Haemoglobin < 5.0g/dl ] ) and 102 ( 27.4% ) had leukocytosis ( WBC > 15,000 / mm3 ) and 4 ( 1.1% ) had leucopaenia. There was no significant difference between Cases and Controls in the prevalence of any of these abnormalities.

82

83

RESULTS OF CEREBROSPINAL FLUID (CSF) ANALYSIS, URINALYSIS AND BLOOD CULTURE

Two hundred and thirty-five children had LP done. This included 105 / 108 Cases and 130 / 265 Controls. LP could not be done on admission in 3 children ( all Cases ), all of whom presented in status epilepticus with a clinical diagnosis of meningitis. All of these had contraindications to lumbar puncture. All the CSF cultures were negative and no organism was seen on gram stain. The mean (SD) of CSF glucose, CSF glucose : Random plasma

84

glucose ratio, CSF protein and CSF white blood cell count of the patients with meningitis compared with that of those without meningitis is shown in Table XIV. All the differences were statistically highly significant ( p <

0.0001 ).

There was no case of bacteraemia or significant bacteriuria. Two (15.4 %) of 13 children with LF had proteinuria and 2 / 13 (15.4%) had haematuria. The prevalence of abnormalities on urinalysis, including microscopic urinalysis is shown in Table XV. The prevalence of glucosuria was higher among the controls ( 4.5% vs 0 ) while the prevalence of proteinuria was higher among the cases ( 11.1% vs 3.0% ). Differences in the prevalence of the other abnormalities were not statistically significant ( p > 0.05 ).

85

86

OUTCOME OF ILLNESS IN CASES AND CONTROLS

Twelve (3.2%) children (3 Cases and 9 Controls) were discharged against medical advice.

None of the 12 had LF. Three hundred and sixty patients ( 105 Cases and 256 Controls ) completed treatment. Four children ( Cases 3, Control 1 ) with LF-malaria co-infection received intravenous ribavirin and intravenous quinine. Nine (3.4%) of the Controls with LF only received intravenous ribavirin. Eighty three children ( 47 Cases, 36 Controls ) diagnosed with malaria only received intravenous quinine. The twenty patients with meningitis ( 10 Cases and 10 Controls ) received intravenous Ceftriaxone.

Thirteen (3.6%) of the 361 children who completed treatment died. Case fatality rate was [ 7 / 105 ( 6.7% ) among the cases and 6 / 256 ( 2.3% ) among the Controls ( OR [ 95% CI ]

87

= 2.99 [ 0.98, 9.12+, Fisher’s exact p = 0.098 ). The 7 deaths among the Cases included one child with LF and six with malaria. The 6 deaths among the controls included two children with LF and one each with malaria , meningitis , diabetic ketoacidosis and myocarditis. The case fatality rate in

LF ( 3 / 13 or 23.1% ) was significantly higher than in those without LF ( 10 / 348 or 2.9% ) ( OR [95% CI] = 10.14 [2.41, 42.59], Fisher’s exact p = 0.017 ).

DISCUSSION

The 2.8% prevalence of LF among children with CAWF was not significantly different from the 3.8% prevalence in the controls. While no previous studies have, to our knowledge, investigated the prevalence of LF in children with CAWF, the results of this study suggest that LF is an important cause of CAWF in children in endemic areas. Even though the prevalence was low compared with that of malaria and meningitis, the importance should derive from the mortality and sequelae associated with LF. ^23,24 With particular reference to children with febrile convulsion, the 1.6% prevalence of LF even though lower than the 4.3% prevalence among children with CAWF excluding febrile convulsion would still be of importance because of the associated significant mortality, sequelae, risk of nosocomial infection and occupational hazard to health workers involved in the care of these children. ^23,24

An earlier study reported a 7.6% prevalence of enteroviruses in Ibadan children with febrile convulsion.¹⁰ Even though the prevalence of LF in febrile convulsions in this study is lower than the prevalence of enteroviruses in the Ibadan study, ¹⁰ the implications of LF in CAWF may be expected to be different from outcomes from other viruses because of the associated mortality and sequelae. Nonetheless, the results of both studies suggest the need and importance for tests to identify viral aetiologic agents in children with CAWF in the tropics, although viral infections are reported to be more associated with febrile convulsions in temperate countries than the tropics. ⁹

The prevalence of CAWF among children admitted to the CHER in this study was 11.8% and is within the range of 3 – 28% reported from earlier studies. 10,11,31,40 This supports the reports

88

from other areas of Nigeria that CAWF is an important emergency in children. 10,11,31,40

The aetiology of fever was unidentified in 25.0% of the children with CAWF in this study.

Unidentified causes of CAWF have similarly been reported in other studies of children with CAWF. 12, 15, 27, 40 Limitations in diagnostic resource is an important factor ⁹³ and the proportion of such children tends to decline as new diagnostic techniques become available. ⁹³ Through the use of LV-RT-PCR, Lassa virus was identified in 2.8% of children presenting with CAWF in the emergency room in this study. It may be a safe assumption that in the absence of this test these children may have been classified with CAWF of unidentified aetiology.

Malaria, with a prevalence of 46.3%, was the commonest cause of CAWF, but not the commonest cause of fever in children without convulsions in this study. This further confirms the leading role of malaria as a cause of fever and CAWF as reported in earlier studies. 5,10,15,31,40 The higher prevalence of malaria in children with CAWF compared with the Control groups of children with fever but no convulsion also supports earlier reports that convulsion in malaria are both febrile convulsions and other CAWF excluding febrile convulsions. ^3,39

The 43.5% prevalence of malaria in children with CAWF in this study, although higher than the 32.7% reported in children with febrile convulsions in Benin City, ³¹ is similar to that in reports from Ibadan (45.2%),¹⁰ but lower than that in reports from Ilorin (71.7%) ⁵ and Jos (74.8%). ¹⁶ In southern Nigeria, malaria transmission is intense all the year round, while in the middle belt and far north the intensity of malarial transmission is more seasonal. 5, 10, 16, 31, 131 This may explain the higher prevalence of malaria in children with CAWF in these areas.

The prevalence of LF was highest in pre-school age children in this study, in keeping with previous reports of a higher prevalence of LF 41, malaria 10, 16, 33, 39, 40 and meningitis 10, 37, 40, 48 in pre-school children. The apparent susceptibility of this age group to several infections, including LF, may not be unrelated to the developmental status of the immune system.

89

The prevalence of LF was similar in children with CAWF (2.8%) and the Controls (3.8 %).

Seizures, confusion and coma are recognised acute neurologic morbidities in LF, ^23-25 but few studies, if any, have in the past prospectively defined the risk of CAWF in LF or the contribution of LF to CAWF. Seizures associated with LF have been reported to be generalized tonic-clonic in nature. ⁴¹ Two (66.7%) of the 3 patients with LF and convulsions in this study had generalised tonic-clonic seizures. A better understanding of the pathogenesis of the neurological manifestations of the Lassa virus may illuminate our knowledge of the interactions between the

virus and the central nervous system. ⁶³

The clinical differentiation of LF from the common causes of fever is difficult. 2, 81-88, 90 The common symptoms of fever, malaise, headache, abdominal pain, sore throat, gastroenteritis and cough 53, 59, 70, 76 were not significantly associated with LF in this study.

The lack of differentiation of LF from the presentation of other common febrile illnesses is consistent with previous reports. 22, 59, 90 Even though the association of an acutely ill appearance with LF approached statistical significance in this study, it may also be a feature in any number of acute febrile illnesses.8, 39, 44, 92, 93 With the exceptions of bleeding and vomiting, no presenting clinical feature had a significant association with the risk of LF as a cause of fever. While vomiting had a significant negative association with the risk of LF, bleeding had a positive association. Even then the sensitivity of the absence of vomiting for LF was only 1/13 or 7.7% while the sensitivity of presentation with bleeding was only 4/13 or 30.8%. A high index of clinical suspicion thus remains important in the diagnosis of LF.

Leucocytosis was found in LF patients in this study. This is similar to findings in previous studies even though it was reported as a less common finding. ^41,

Plasmodium falciparum was the only specie of malaria parasites identified in the present study. This is similar to the findings of other workers. 7, 13, 16, 40 Although researchers at Ibadan ³⁹ reported Plasmodium malariae in addition to Plasmodium falciparum, the proportion of those with Plasmodium malariae was less than 3%. ³⁹ The absence of Plasmodium malariae in more recent studies, 7, 13, 16, 31 as well as in this study may be due to the effect of pre-hospital treatment with anti-malarial drugs.

In document El cuerpo del delito (página 125-129)