Skip to main content

Table 3 Univariable and multivariable penalised logistic regression on the potential prognostic factors for maternal death

From: Severe falciparum malaria in pregnancy in Southeast Asia: a multi-centre retrospective cohort study

 

Univariable

Multivariable

Characteristic

N

Odds ratio (95% CI)

p-value

Odds ratio (95% CI)

p-value

Age < 20 years

6/42

3.75 (0.95–14.90)

0.06

  

 20–29

3/83

Reference

   

  ≥ 30

7/54

3.21 (0.85–12.08)

0.08

  

Trimester First

2/22

0.94 (0.19–4.61)

0.94

  

 Second

7/71

Reference

   

 Third

8/73

1.13 (0.38–3.34)

0.82

  

 Postpartum

1/3

6.96 (0.76–63.66)

0.09

  

Gravidity primigravida

4/46

Reference

   

 Multigravida

6/77

0.86 (0.24–3.03)

0.81

  

Presence of gametocytes

4/40

3.69 (0.71–19.28)

0.12

  

 No

2/90

Reference

   

Presence of schizont

1/17

1.93 (0.28–13.16)

0.50

  

 No

4/99

Reference

   

Presence of pigmenta

7/103

2.41 (0.12–48.57)

0.57

  

 No

0/29

Reference

   

Fever > 37.5 °C

9/99

0.89 (0.29–2.76)

0.84

  

 No

7/84

Reference

   

Coma

22/70

11.40 (3.56–36.50)

 < 0.0001

7.18 (2.01–25.57)

0.002

 No

4/141

Reference

 

Reference

 

Severe anaemia

6/69

1.44 (0.47–4.47)

0.52

  

 No

14/132

Reference

   

Renal failure

5/17

4.79 (1.21–18.99)

0.026

  

 No

13/171

Reference

   

Respiratory failure

5/20

12.95 (3.16–53.12)

0.0004

4.98 (1.13–22.01)

0.03

 No

16/171

Reference

 

Reference

 

Convulsion

1/7

2.70 (0.38–19.31)

0.32

  

 No

21/192

Reference

   

Metabolic acidosisb

7/35

2.10 (0.54–8.16)

0.28

  

 No

9/55

Reference

   

Jaundice

5/30

1.37 (0.40–4.70)

0.62

  

 No

16/172

Reference

   

Hypoglycaemia

3/12

3.66 (0.79–16.95)

0.10

  

 No

17/137

Reference

   

Hypotension

2/6

10.83 (1.84–63.89)

0.009

11.21 (1.27–98.92)

0.03

 No

18/192

Reference

 

Reference

 

Hyperparasitaemia

8/86

1.40 (0.49–4.02)

0.53

  

 No

13/120

Reference

   

Prostration

4/18

2.74 (0.39–19.19)

0.31

  

 No

4/98

Reference

   
  1. CI confidence interval
  2. Penalised logistic regression with fixed-effects for each study site was used. Multiple imputation was used for multivariable models. Calendar year and treatment (quinine-based or artemisinin-based) were adjusted in multivariable models
  3. aWhen only patients with information on malaria pigment quantitatively assessed, the unadjusted odds ratio of maternal deaths for women with pigment in neutrophil (> 5%) was 2.33 (3/22 vs 2/29, 95%CI 0.34–15.83, p = 0.39)
  4. bVariables not assessed for multivariable model because missing > 50%