Shobowale EO, Solarin AU, Elikwu CJ, Onyedibe KI, Akinola IJ, Faniran AA. Neonatal sepsis in a Nigerian private tertiary hospital: Bacterial isolates, risk factors, and antibiotic susceptibility patterns. Ann Afr Med 2017;16:52-8.

Sir,

Shobowale et al.[1] in their article on neonatal sepsis, reported about findings suggesting that the prevalence of neonatal sepsis was higher in “out born” neonates, those with prematurity, premature rupture of membranes (PROMs) and those aged <3 days. There are, however, up to 17 notably contentious issues here outlined that cast some doubt about the validity of their findings, which I seek your kind indulgence Sir, to publish these for the authors to clarify.

The abstract stated a study duration as “a 1-year period.” but the methodology stated “from August 2014 to August 2015”, which is a 13-month period

Inclusion criteria: The statement that “All infants satisfying the criteria for sepsis will be subjected to blood culture” suggests a study proposal statement, not what should be used for a completed study being reported

Only blood samples were chosen for analysis; other samples may yieid culture positive when results for blood culture are negative, especially where antibiotics have been previously administered before presentation. Furthermore, standard protocol for sepsis evaluation is inclusive of other samples such as cerebrospinal fluid, urine, stools, aspirates, wound, and mucosal swabs[23]

The researchers did not address the possible compounding effect of prior antibiotic use for the subjects or their mothers before presentation to hospital

It was not clear whether just one or all of the criteria for sepsis had to be present or a combination of a number of them occurring for a clinical diagnosis of neonatal sepsis to be made

The statement that “parents who refuse to provide informed consent shall be excluded…….” is suggestive of a study proposal statement

There is no statement on whether ethical approval for the study was sought for or obtained from an appropriate body

A certain formula (“Kish;” with no reference of its source), was used to obtain a sample size of 120; but only the results of 100 neonates were analyzed, and no account was made of the difference

There is a disparity between the list of antibiotics tested for sensitivity as stated in the methodology and what was displayed as shown in Figure 1 of this letter (captioned as Figure 2 in Shobowale, et al).[1] The list included ampicillin, vancomycin, erythromycin, gentamicin, and oxacillin, but these were not represented on the chart [Figure 1]; rather strangely, the chart has bars for 6 other antibiotics that were not listed in the methodology (clindamycin, piperacillin/Tazobactam, cefuroxime, colistin, ampicillin/sulbactam, and cefepime)

Figure 1

Antimicrobial susceptibility profiles of pathogens isolated in neonatal sepsis

Results:

The last sentence of the opening paragraph of “Results” stated that “Sepsis was equal in incidence in both early and late-onset variants [” However, the table has no data to support this statement, and it is not clear where the basis for this statement came about

It was stated that “majority of parents were in the low socioeconomic bracket, n = 48 (38.9%),” but this is contrary to what is in Table 2, where 52 were majority (high 16, plus middle 36); the percentages were also incorrect

Table 2

Outcome of neonates with bacteriologic sepsis

Variable	Outcome		OR	95% CI	χ2	P
	Died	Discharged
Age (days)
<3	8 (12.1)	58 (87.9)	1.04	0.45-5.1	0.003	0.96
>3	4 (11.8)	30 (88.2)
Birth weight (g)
>2500	4 (5.6)	68 (94.4)	0.15	0.04-0.6	10.11	0.004
<2500	8 (28.6)	20 (71.4)
Birth status
Preterm	8 (21.1)	30 (78.9)	3.87	1.1-13.9	4.76	0.036
Term	4 (6.5)	58 (93.5)
PROM
Present	6 (46.2)	13 (53.8)	5.77	1.6-20.7	8.52	0.007
Absent	6 (10.5)	75 (89.5)
Parity
<3	8 (10.5)	68 (89.5)	0.62	0.17-2.3	0.99	0.32
>3	4 (16.7)	20 (83.3)
Delivery
BUTH	8 (42.1)	13 (57.9)	5.85	1.6-20.9	2.71	0.007
Sepsis
Outside BUTH	4 (5.1)	75 (74.9)	0.35	0.08-1.5	1.41	0.16
Early onset	9 (53)	8 (47)
Late onset	13 (76)	4 (34)

OR=Odds ratio, CI=Confidence interval, PROM=Premature rupture of membranes, BUTH=Babcock University Teaching Hospital

It was also stated that “Ten of the neonates had preterm PROMs (10%).” It is not clear what “preterm PROMs” mean; the term does not represent a recognized entity.

It is not clear why the percentages of some number of cases in Table 1 contained decimals (e.g., categories of “Growth” were 34 [34.3%] and 66 [65.7%]), and why some number of cases do not round up to 100 (e.g., categories under “Gestation” had Preterm as 19 and Term as 35, which rounds up to 54). Percentages for “Birth weight” categories (>2500 g and <2500 g), socioeconomic status and Gender were also wrong. Since there were 100 cases, the percentages of any number out of these 100 cases should correspond to the numbers with no decimal places

Table 1

Summary statistics of neonates

Variable	n (%)
Age (days)
<3	65 (65)
>3	35 (35)
Growth
Positive	34 (34.3)
Negative	66 (65.7)
Outcome
Died	12 (12)
Discharged	88 (88)
Gestation
Preterm	19 (35.2)
Term	35 (64.8)
Birth weight
>2500	72 (42.5)
<2500	28 (57.5)
Socio-economic status
High	16 (14.9)
Middle	36 (36.2)
Low	48 (38.9)
Meconium stained amniotic fluid
Present	2 (2)
Absent	98 (98)
Premature rupture of membranes
Present	12 (12)
Absent	88 (88)
Prolonged labour
Yes	17 (17)
No	83 (83)
Gender
Male	52 (52.5)
Female	48 (47.5)

The categorizations for age (<3 and > 3 days) did not capture age of exactly 3 days at presentation; similarly for “Birth weight,” weight of exactly 2500 g was not included

In Table 2, the figures and percentages for “Early” and “Late” onset sepsis do not tally for both “Died” and “Discharged” categories

In spite of much higher proportionate mortality among neonates born “Within” the hospital than neonates in the “Out born” category (8 out of 21 or 38.1%, compared to 4 out of 79 or 5.1%), yet the authors stated that delivery outside a tertiary hospital was a risk factor for mortality. This is a total misinterpretation

It was stated that “The risk factors for sepsis in this study was age <3 days and prematurity;” but Table 3 shows that proportionately more neonates aged >3 days were culture positive compared to those <3 days (17 out of 35 or 48.6% compared to 17 out of 65 or 26.2%). Another misinterpretation

Table 3

Risk factors for neonatal sepsis

Variable	Growth		OR	95% CI	χ2	P
	Negative	Positive
Age (days)
<3	48 (73.8)	17 (26.2)	2.61	1.1-6.2	5.09	0.03
>3	18 (51.4)	17 (48.6)
Anaemia
Yes	4 (57.1)	3 (42.9)	0.69	0.15-3.3	0.21	0.65
No	61 (65.6)	32 (34.4)
Birth status
Preterm	42 (73.7)	15 (26.3)	17.3	6.1-49.1	35.03	0.001
Term	6 (13.9)	37 (86.1)
Booking status
In born	39 (62.9)	23 (37.1)	1.06	0.46-2.43	0.69	0.89
Out born	27 (71.1)	11 (28.9)
Gender
Female	32 (66.7)	16 (33.3)	0.72	0.31-1.65	0.62	0.43
Male	34 (65.4)	18 (34.6)
Prolonged labour
Yes	11 (64.7)	6 (35.3)	0.93	0.31-2.79	0.02	0.91
No	55 (66.2)	28 (33.8)
PROM
Present	6 (50)	6 (50)	0.44	0.13-1.49	1.78	0.19
Absent	61 (69.3)	27 (31.7)

OR=Odds ratio, CI=Confidence interval, PROM=Premature rupture of membranes

The data for “Birth status” in Table 3 did not seem to tally correctly; there were 34 “Positive” cases of bacterial isolation, but the table has 15 “Preterm” and 37 “Term” neonates, which gives 52 cases rather than 34. Similarly, the data for “PROM” do not tally correctly to give total of 34 positive bacterial cultures

Table 3 has “Booking status” but with categories as “In born” and “Out born;” booking status is about antenatal care and not whether a neonate is born in or out of the hospital. Besides, there is no place in the article where antenatal care information was captured in the data for risk factors.

CONCLUSIONS

There are serious methodological flaws and inconsistencies that need to be clarified to make the findings valid

There are overt disparities and inconsistencies in antibiotic testing and sensitivity results; similarly, the data analysis has lots of calculation errors and misinterpretations resulting in incorrect statements on the conclusions

If the results could be considered based on some of the findings, it seems that with a higher rate of positive blood cultures among neonates >3 days old and higher mortality among inborn neonates, it is possible that nosocomial infection was a significant contributor to this mortality rate in this hospital. This is especially likely with the higher proportion of positive blood cultures found among neonates aged >3 days (late-onset neonatal sepsis), and with isolation of such organisms as Klebsiella pneumoniae, Proteus mirabilis, Staphylococcus aureus plus Coagulase negative staphylococcus (CONS), which are recognized agents known to cause hospital-acquired infection[3456]

The validity of findings of this study can only be ascertained if the errors and inconsistencies could be clarified, along with correct interpretation of these findings.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.