Types of Bacteria associated with Neonatal Sepsis in Al-Thawra University Hospital, Sana’a, Yemen, and their Antimicrobial Profile

Abstract

Objectives:

This study was undertaken to investigate the organisms causing sepsis in the Neonatal Unit at Al-Thawra Hospital, Sana’a, Yemen, determine their resistance to antibiotics, and recommend policy for empirical treatment.

Methods:

A total of 158 neonates having one or more signs of sepsis, and aged from 0 to 28 days, were enrolled in this study. A blood sample was taken from each subject, cultured, and then antibacterial susceptibility tests were performed for isolates.

Results:

90 (57%) cases yielded positive cultures. Early-onset sepsis showed higher positive culture results (61.7%) than late-onset sepsis (32%). Significant positive culture results were found among the group with birth weight 0.9–2 Kg (78.6%). Gram negative bacteria constituted 97.8% of the total isolates, of which Klebsiella pneumoniae was the predominant pathogen (36.7%), followed by Pseudomonas species (30.0%). The commonest symptoms among the cases were respiratory distress (72.2%), jaundice (62.2%), cyanosis (51.1%), and lethargy (47.8%); the mortality rate was 27.8%. All Gram negative bacterial isolates were sensitive to imipenem and some isolates were sensitive to fourth-generation cephalosporins, but most isolates were highly resistant to the majority of other antibiotics tested.

Conclusion:

Gram negative organisms were the most frequent causative agents of bacterial sepsis, which is a significant cause of mortality and morbidity in the newborn, and particularly in those of very low birth weight. It can also be concluded that imipenem and fourth-generation cephalosporins can be used for empirical treatment of bacterial sepsis.

Advances in knowledge

1. This study provides current data on the organisms causing sepsis in neonatal units and the resistance of the sepsis-causing organisms to the relevant antibacterial drugs. It was found that Gram negative organisms were the most frequent causative agents of bacterial sepsis (97.8%).

2. This information is important in order to recommend policy for empirical treatment.

3. This is one of only a few studies from Arabic countries focused on neonatal sepsis.

Application to patient care

1. The research reported here could aid in the prevention of neonatal sepsis and build a treatment policy for neonatal bacterial sepsis.

2. This study alerts physicians to the defective hygiene in neonatal units and the high rate of resistance to antibiotics.

Sepsis is a significant cause of mortality and morbidity in the newborn, particularly among those of very low birth weight and premature infants.1,2 The World Health Organization (WHO) estimates that worldwide 1.6 million newborn babies die every year from neonatal infections. Individual units have reported case fatality rates as low as 2–3% and as high as 50%,3 but the average overall mortality from neonatal sepsis reported from many neonatal intensive care units in developed countries is around 20%.4 Despite recent advances in neonatal intensive care and current strategies to treat neonatal sepsis, mortality rates have not fallen for over three decades4 except in babies born to mothers who have received intrapartum prophylaxis (IAP) for Group B Streptococcus (GBS).5 Whilst IAP has been successful in significantly reducing the incidence of early onset GBS disease in the newborn it has led to an increase in Gram negative infections in institutions where antibiotics other than penicillin are used for IAP.6 Worryingly, both short and long term neurodevelopmental morbidity in survivors of neonatal sepsis is also significant.7 The study by Stoll et al. in 2002 showed that neonatal infection is associated with a 30–80% increase in neurodevelopmental impairment and a 30–100% increase in odds for poor head growth which is a very good predictor of long-term morbidity.6

This study aimed to determine the types of bacteria associated with neonatal sepsis in neonatal unit of Al-Thawra University Hospital in Sana’a, Yemen, and their antibiotic profile.

Methods

This prospective study was carried out during a 12 month period from April 2008 to March 2009. The study proposal was approved by the Department of Medical Microbiology, in the Faculty of Medicine & Health Sciences at Sana’a University, Yemen. A total of 158 hospital-born neonates with a clinical diagnosis of septicaemia, and admitted to the neonatal unit of Al-Thawra University Hospital in Sana’a city, were included in this study. A consent form was completed by the parents for each subject.

Early-onset sepsis (EOS) was defined as sepsis occurring within the first week of life. Late-onset sepsis (LOS) was defined as sepsis that occurred any time after the first week of life until the end of the neonatal period (28 days of age). A full history was taken from doctors and files, and the findings were recorded in a predesigned questionnaire. The data collected included: any complications during delivery, characteristics of the infant, sex, birth weight, temperature, other medical conditions in the infant, major signs and symptoms and other necessary data for the research. Blood culture was preformed for all neonates suspected of having septicaemia. Septicaemia was suspected in the following settings: 1) At birth: All neonates born to mothers with maternal fever, prolonged rupture of membranes for more than 24 hours, or foul-smelling or meconium-stained liquor, and those having severe prematurity or birth asphyxia necessitating active resuscitation; 2) After birth: All newborns with lethargy, poor feeding, abdominal distension, respiratory distress, instability in temperature, pathological jaundice, and convulsions.

The collected data and the results of this study were statistically analysed for means of significance by the Epi Info computer programme, Version 6 (Epi Info, Centers for Disease Control (CDC), Atlanta, Georgia, USA). The specimens were collected as follows: one blood culture at least was obtained for each neonate with signs of sepsis, by the method described by Buttery.8 Blood from peripheral venous or arterial punctures was collected under sterile conditions as described by Hall and Lyman.9 The blood specimen was then inoculated into a BACTEC vial (BACTEC, Rochester, UK) and the inoculated culture bottles were incubated as soon as possible in the BACTEC brand fluorescent series instrument (BACTEC 9120) for up to 5 days, as recommended by Becton Dickinson Microbiological Systems.10 Positive blood culture vials were sub-cultured into blood agar, chocolate agar, MacConkey agar, and Sabouraud agar (for yeast), then the media were inspected for the presence of bacterial growth. The isolated bacteria were defined by colonial morphology, Gram stain reaction, biochemical reactions and confirmed by API-20E biotyping (API 20E system, Analytab Products, Plainview, New York, USA). Finally, the sensitivity of isolated bacteria to various antibiotics was assessed according to the modified Kirby and Bauer method on Mueller-Hinton agar. The concentrations of the antibiotics used (μg/disc) were as follows: ampicillin (AMP) − 10; gentamycin (CN) − 10; amikacin (AK) − 30; ceftazidime (CAZ) − 30; ceftriaxone (CRO) − 30; cefepime (FEP) − 30; piperacillin (P) − 100; imipenem (IPM) −10; and aztreonam (ATM) − 30. The Clinical and Laboratory Standards Institute (CLSI) guidelines were used to determine antimicrobial resistance.

Results

This study included 158 neonates, their ages ranging from 0 to 28 days with a mean age of 5.2 days.

The standard deviation (SD) was equal to 6.1 days, mode equal to 1 day and median equal to 3 days. There were 96 (60.8%) males and 62 (39.2%) females. The detailed results of this study are presented in Tables 1 to 6. Table 1 shows the distribution of positive bacterial culture according to sex. A total of 90 (57%) of patients had a positive culture for bacterial growth. The rate of positive culture among females (62.9%) was higher compared to males (53.1%), but the variation was not statistically significant in which χ² = 1.47 and P = 0.22. The EOS showed higher positive culture results (61.7%) than LOS (32%), with the relative risk of contracting EOS bacterial infection 2.2 times that of LOS, but the difference was not statistically significant in which χ² = 2.7 and P = 0.1 [Table 2]. Table 3 shows that highly significant positive cultures results were found among the group with birth weight 0.9–2 Kg (78.6%, χ² = 5.7, P = 0.01). The commonest symptoms among the cases studied were respiratory distress (72.2%), jaundice (62.2%), cyanosis (51.1%) and lethargy (47.8%). On the other hand, lethargy, apnoea, poor feeding and unconsciousness occurred significantly less frequently among positive bacterial culture cases in which χ² > 3.7 and P <0.05. However, the mortality rate was higher among positive culture cases (27.8%), than negative ones (2.9%) and the variation was statistically significant in which χ² = 16.86 and P = 0.0004 [Table 4].

Table 1:

The rate of positive bacterial culture among male and female neonates.

Sex:	Total cases		Cases with Positive Culture		χ2	P
	No.	%	No.	%
Male	96	60.8	51	53.1	1.47	0.22
Female	62	39.2	39	62.9
Crude rate	158	100.0	90	57

Legend: χ² = chi square; P = probability value.

Note: χ² > 3.7 and P <0.05 are significant.

Table 6:

Resistant pattern of the isolates to some antibiotics

Antibiotic	Klebsiella spp (n = 33)	Pseudomonas spp (n = 27)	Enterobacter spp (n = 13)	Citrobacter spp (n = 4)	Acinetobacter spp (n = 10)
AMP	29 (87.8)	23 (85.2)	12(92.3)	3 (75.0)	5 (50)
AML	31 (93.9)	19 (70.3)	13 (100)	4 (100)	9 (90)
ATM	30 (90.9)	20 (74.0)	10(76.9)	4 (100)	7 (70)
PRL	29 (87.8)	3 (11.1)	8 (61.5)	1 (25.0)	1 (10)
CAZ	30 (90.9)	10 (37.0)	12(92.3)	4 (100)	5 (50)
FEP	12 (36.4)	4 (14.8)	0	1 (25.0)	0
CRO	32 (96.9)	19 (70.3)	13 (100)	4 (100)	9 (90)
IPM	0	0	0	0	0
AK	13 (39.3)	12 (44.4)	8 (61.5)	0	5 (50)
CN	9 (27.2)	17 (62.9)	5 (38.4)	0	5 (50)

Legend: AMP = ampicillin; AML = amoxicillin; ATM = azteronam; PRL = pipracillin; AK = amikacin; CN = gentamicin; CAZ = ceftazidim; CRO = ceftriaxon; FEP = cefipime = c; IPM = imipenem.

Table 2:

The rate of positive bacterial culture in early-onset (EOS) and late-onset of sepsis (LOS) and their relative risks for contracting bacteria, confidence interval and test of significance

Type of sepsis	Positive Bacterial Culture		RR	CI	χ2	P
	No.	%
EOS ≥ 7 days n =133	82	61.7	2.2	(0.77–6.3)	2.7	0.10
LOS < 7 days n = 25	8	32	0.7	(0.4–1.2)	2.7	0.10
Total	90	57

Legend: RR = Relative risk; CI = Confidence Interval; χ² = chi square, P = probability value

Note: χ² > 3.7 and P < 0.05 are significant

Table 3:

The rate of positive bacterial culture and its relationship to birth weight

Birth weight (Kg)	Total cases		Cases with Positive culture		χ2	P
	No.	%	No.	%
0.9–2	73	46.2	50	68.5	5.7	0.01
2.1–3	63	39.9	31	49.2	3.7	0.05
3.1–4	19	12	08	42.1	0.2	0.68
<4	3	1.9	1	33.3	0.7	0.4

Legend: χ² = chi square; P = probability value.

Note: χ² > 3.7 and P <0.05 are significant

Table 4:

Presenting clinical symptoms and mortality rate among total patients compared with positive bacterial culture.

Symptoms	Negative culture n = 68		Positive culture n = 90		χ2	P
	No.	%	No.	%
Fever, n = 20	8	11.8	12	13.3	0.09	0.76
Lethargy, n = 64	21	31	43	47.8	4.6	0.03
Jaundice, n = 98	36	53	56	62.2	0.00	0.95
Apnoea, n = 15	3	4.4	12	13.3	3.6	0.05
Respiratory distress, n = 116	51	75	65	72.2	0.15	0.7
Poor feeding, n = 19	2	2.9	17	18.9	9.3	0.002
Cyanosis, n = 76	30	44.1	46	51.1	0.76	0.38
Abdominal distention, n = 7	4	5.8	3	3.3	0.6	0.4
Pallor, n = 20	5	7.3	15	16.7	3.04	0.08
Convulsion, n = 5	1	1.5	4	4.4	1.12	0.29
Birth asphyxia, n = 8	4	5.9	4	4.4	0.17	0.68
Unconscious, n = 21	5	7.3	16	17.8	3.65	0.05
Mortality rate, n = 27	2	2.9	25	27.8	16.86	0.0004

Legend: χ² = chi square; P = probability value.

Note: χ² > 3.7 and P <0.05 are significant

Gram negative bacteria constituted 97.8% of the total isolates, of which Klebsiella pneumoniae was the predominant pathogen with 36.7%, followed by Pseudomonas species with 30.0%. Citrobacter species was the least common isolate of Gram negative bacteria with a percentage of 4.4%, while Gram positive bacteria rates were very low (1.1%) [Table 5]. All Gram negative bacterial isolates were sensitive to imipenem and some isolates were sensitive to fourth-generation cephalosporins (cefepime), but most isolates were highly resistant to the majority of other antibiotics tested [Table 6].

Table 5:

The frequency and percentage of bacterial species that isolated from neonate sepsis cases

Organism	No.	%
Gram negative bacteria	88	97.8
Klebsiella pneumoniae	33	36.7
Pseudomonas species	27	30
Enterobacter species	14	15.6
Acinetobacter species	10	11.1
Citrobacter species	4	4.4
Gram positive bacteria	1	1.1
Staphylococcus aureus	1	1.1
Fungi	1	1.1
Candidaspecies	1	1.1
Total positive bacteria	90	100.0

Discussion

In this study, blood culture obtained from 158 neonates yielded positive results in 57% of cases, while the remaining 43% gave negative results. A similar result was reported by Ako-Nai and others in Nigeria in 1999 with a positive percentage of 55%.11 Slightly higher positive results were reported by Rohsiswatmo (65.3%) in Indonesia,12 Rahman et al.13 in Pakistan (62.8%) and by Macharashvili et al. 14 in Georgia (63%). In contrast, much lower positive results were reported from Iran (5.6%),15 Kuwait (8.7%)16 and Saudi Arabia (5%).17 These variations can be attributed to many different factors of which antibiotic therapy prior to the laboratory diagnosis may have had the most important influence on the low culture results.

The EOS rates (61%) were higher than the LOS rates (32%) in this study. This result is compatible with the reports from Iran and Pakistan,18,19 but it contrasts with a report from Bangladesh, where LOS was more common than EOS.20 However, the most common organism in EOS was K. pneumoniae; the same finding was recorded in many developing countries.21–23 In the present study, females showed slightly higher positive results than males, which is similar to findings reported from Iran.24

The commonest symptoms among the cases studied were respiratory distress (72.2%), jaundice (62.2%), cyanosis (51.1%, and lethargy (47.8%). On the other hand, lethargy, apnoea, poor feeding and unconsciousness showed significantly low rates of occurrence among positive bacterial culture cases. In general, the presenting clinical symptoms and signs in our study agreed with the WHO clinical criteria for neonatal sepsis.25

There was a significant correlation between low birth weight and the incidence of neonatal sepsis with, statistically significant values of χ² = 5.7 and a P value equal to 0.01. The same finding was observed in the studies of Al-Umran and Twum-Danso in Saudi Arabia and Jeong in Korea.17,26 Mortality rate among positive culture cases in this study was 27.8%, which was similar to that found in Zimbabwe,27 Pakistan28 and Georgia.14

The most common organisms which caused EOS and LOS in the present study were Gram negative bacteria, with a percentage of 97.8%. This result agreed with other related studies in many developing countries,13,14,23,29 but differed from those reported by Huda et al.16 in Kuwait, Haque et al.4 in the UK and Hyde et al.30 in the USA where Gram positive bacteria were more predominant. K. pneumoniae was the predominant isolate in the present study, representing 36.7% of the total isolates. Similar findings were reported by Macharashvili in Georgia,14, Mehdinejad et al. in Iran (33.5%)15 and by Kumhar et al. in India (33.8%),21 but lower results were reported from Nepal (18.3%) and Pakistan (7.6%).31,13 Moreover, it was found in this study that Pseudomonas species were the second most common isolate; similar findings were reported in Pakistan.13,32 The overwhelming majority of Gram negative organisms isolated in this study and among the other studies reviewed suggests that these infections may in fact be acquired from the hospital or community environment, due to poor hygienic practices during delivery and postnatal care, rather than reflecting vertical transmission to the infant from exposure to vaginal tract flora.33,34 Another possible explanation of the predominance of Gram negative bacteria in our study is that asymptomatic colonised patients, the contaminated environment, or both can serve as reservoirs for these pathogens, which are then spread by the hands of health care workers.35,36 Furthermore, a limitation of our study was that there is insufficient sterility testing or wipe testing of the neonatal wards to confirm if the conditions are unhygienic which could mean that contaminating bacteria might also be recovered from the environment.

Multi-drug resistance of the causative organisms of sepsis is a rapidly emerging and potentially disastrous problem. Infection with resistant organisms has been associated with treatment failure, higher morbidity and mortality and increased costs. The present study shows a very high degree of resistance of Gram negative organisms to commonly used antibiotics (AMP and CN) as well as third-generation cephalosporins. These results agree with many other studies.12,18,37,,38 The high burden of infection and the degree of antibiotic resistance seen in this study population may signify an emerging trend among sepsis isolates recovered from neonates in developing countries. In a review of data from developing countries, Zaidi et al. found rates of neonatal infection in hospitalised infants to be 3–20 times higher than those in developed countries, and approximately 70% of these infections would not be susceptible to conventional empirical antibiotic regimens, such as AMP with CN.39

All Gram negative organisms in this study were sensitive to IMP; the same results were reported by Shaw et al., Al-Tawfiq and Antony, Rohsiswatmo, and Waseem et al.31,40,12,23 In contrast, a study in the Philippines reported 20% resistance to IMP.36 In the present study, there was increasing resistance to third-generation cephalosporins, probably attributable to extended spectrum beta-lactamase (ESBL) production by Gram negative bacteria, especially Klebsiella, as also observed in another studies on neonates. 31,39,41 The lack of culture-driven antimicrobial therapy and limited, consistent infection control practices is likely to be responsible for the resistance of Gram negative organisms in this study.

Conclusion

Gram negative organisms were the most common causative agents of bacterial sepsis in this study. In turn, they are a significant cause of mortality and morbidity in the newborn, particularly in very low birth weight infants. IPM and fourth-generation cephalosporins such as FEP can be used for the empirical treatment of bacterial sepsis in Sana’a city, Yemen.