Lower antibiotic prescription rates in hospitalized COVID-19 patients than influenza patients, a prospective study

Figures & data

Figure 1. Study design. Local influenza and COVID-19 patients were included from Haukeland University Hospital and Haraldsplass Deaconess Hospital during the 2018/2019 influenza season and spring/summer of 2020. The national cohort included COVID-19 patient data from the Norwegian Intensive Care and Pandemic Registry. Inclusion criteria were age > = 18 years, and a diagnosis of either influenza in 2018/2019 or COVID-19 in 2020/2021.

Table 1. Clinical characteristics and outcomes of COVID-19 and influenza patients.

	Cohorts			p-values
	Influenza	Local	National	Influenza versus	Influenza	Local
	n = 215	COVID-19	COVID-19	Local	Versus	COVID-19 versus national COVID-19^d
		n = 82	n = 2300	COVID-19	National COVID-19
Demographics
Age (median, IQR)^a	65 (47–78)	57 (45–72)	61 (48–74)	.048*	.168	.083
Gender (male)	109/215 (51%)	44/82 (54%)	1362/2300 (59%)	.648	.015*	.282
BMI (median, IQR)^a	24.8 (22.5–28.3)	27.4 (24.6–31.6)	27.4 (24.7–31.4)	<.001*	<.001*	.655
Smoker	34/212 (16%)	5/82 (6%)	72/2300 (3%)	.027*	<.001*	.139
Time from symptoms to admission (median days, IQR)^a	3 (1–4)	7 (4.3–10)		<.001*
Temperature > 37.5 °C	188/215 (87%)	61/82 (74%)	216/378 (57%)	.006*	<.001*	.005*
Pregnant	3/102 (3%)	0	13/938 (1%)		.053
Comorbidities (any)	170/214 (79%)	58/82 (71%)	1516/2287 (66%)	.111	<.001*	.409
Diabetes	30/215 (14%)	8/82 (10%)	400/2300 (17%)	.333	.200	.065
Chronic lung disease (including asthma and COPD)	81/215 (38%)	23/82 (28%)	514/2300 (22%)	.120	<.001*	.209
Cardiovascular disease (including hypertension)	102/215 (47%)	26/82 (32%)	935/2300 (40%)	.014*	.053	.093
Chronic renal disease	28/215 (13%)	5/82 (6%)	134/2300 (6%)	.090	<.001*	.923
Chronic hepatic disease	1/215 (1%)	3/82 (4%)	30/2300 (1%)	.065^e	.286	.064
Chronic neurological disease	52/215 (24%)	7/82 (9%)	101/2300 (4%)	.003*	<.001*	.076
Cancer	18/215 (8%)	5/82 (6%)	105/2300 (5%)	.512	.013*	.494
Immunosuppression	24/215 (11%)	8/82 (10%)	112/2300 (5%)	.727	<.001*	.031*
First antibiotic prescription before 24 h of admission
Any antibiotics	142/215 (66%)	36/82 (44%)	824/2188 (38%)	<.001*	<.001*	.252
Penicillin	123/215 (57%)	27/42 (33%)	384/2188 (18%)	<.001*	<.001*	<.001*
Penicillin with ß-lactamase inhibitor	8/215 (4%)	3/82 (4%)	84/2188 (4%)	1.000^e	.931	1.000^e
Aminoglycosides	44/215 (20%)	11/82 (13%)	81/2188 (4%)	.162	<.001*	<.001*
Cephalosporins	31/215 (14%)	10/82 (12%)	309/2188 (13%)	.619	.905	.621
Fluoroquinolones	2/215 (1%)	0/82 (<1%)	57/2188 (3%)	1.000^e	.165^e	.268^e
Carbapenems	1/215 (<1%)	0/82 (<1%)	5/2188 (<1%)	1.000^e	.431^e	1.000^e
Macrolides	9/215 (4%)	2/82 (2%)	33/2188 (2%)	.733^e	.004*	.363^e
Other antibiotics in	21/215 (10%)	2/82 (5.6%)	47/2188 (2%)	.049*^e	<.001*	.697^e
First antibiotic prescription after 24 h of admission
Any antibiotics	6/215 (3%)	4/82 (5%)	327/2188 (15%)	.471^e	<.001*	.010*^e
Narrow-spectrum^b	3/215 (1%)	1/82 (1%)	148/2188 (7%)	1.000^e	<.001*^e	.040*^e
Broad spectrum^c	3/215 (1%)	3/82 (3%)	220/2188 (10%)	.352^e	<.001*^e	.058^e
Clinical outcomes
Chest X-ray infiltrates	63/181 (35%)	57/78 (73%)	1328/1983 (67%)	<.001*	<.001*	.260
Length-of-stay (median days, IQR)^a	2.0 (1.0–5.0)	5.0 (2.0–8.0)	5.0 (2.4–9.5)	<.001*	<.001*	<.001*
30 day mortality	4/215 (2%)	3/82 (4%)	160/2300 (7%)	.399	.002*	.369

^aMann–Whitney U test.

^bPenicillins without penicillinase-activity and aminoglycosides.

^cCarbapenems, cephalosporins, macrolides, tetracyclines, quinolones, piperacillin/tazobactam and others.

^dOverlapping patients (n = 76) were excluded from analysis.

^eFisher’s exact test.

*p-value <.05. p-values < = 0.05 were considered significant.

Chi-square statistics were used unless otherwise noted.

Figure 2. Monthly COVID-19 hospital admissions and antibiotic prescriptions from February 2020 to January 2021. Upper part: National COVID-19 hospital admissions per month (green line). Admissions peaked during spring and autumn of 2020 corresponding to the first and second pandemic wave (divided by the vertical dotted line). Lower part: Proportion of admitted patients receiving antibiotics any time during admission (pink line) and within 24 h of admission (purple line).

Figure 3. Adjusted odds ratios for antibiotic prescription. Adjusted odds ratios (aOR) for antibiotic prescription in (a) COVID-19 patients compared to influenza patients, (b) local COVID-19 patients compared to national COVID-19 patients in the first pandemic wave and (c) national COVID-19 patients in the second compared to first pandemic wave. Odds were adjusted for chest X-ray infiltrates, age and comorbidities.

Table 2. Clinical characteristics of COVID-19 patients during the first and second pandemic wave.

Demographics	First wave n = 1059	Second wave n = 1129	Odds ratio (95% CI)	p-value
Age (median, IQR)^a	60 (49–73)	60 (47–74)		.892
BMI (median, IQR)^a	27 (24–30)	28 (25–32)		.002*
Length-of-stay (median days, IQR)^a	5.8 (2.8–11.1)	4.6 (2.1–8.0)		<.001*
Gender (female)	432 (41%)	464 (41%)	0.99 (0.85–1.18)	.992
Known comorbidity	654 (62%)	781 (69%)	1.39 (1.16–1.66)	<.001*
Diabetes	146 (14%)	229 (20%)	1.61 (1.29–2.01)	<.001*
Chronic lung disease	216 (21%)	268 (24%)	1.22 (0.99–1.49)	.060
Chronic heart disease	384(36%)	491 (44%)	1.35 (1.14–1.60)	.001*
Chronic renal disease	56 (5%)	64 (6%)	1.08 (0.74–1.56)	.696
Chronic hepatic disease	13 (1%)	16 (1%)	1.16 (0.55–2.42)	.699
Chronic neurological disease	47 (5%)	45 (4%)	0.98 (0.66–1.46)	.928
Cancer	48 (5%)	50 (4%)	0.98 (0.65–1.46)	.907
Immunosuppression	60 (6%)	41 (4%)	0.63 (0.42–0.94)	.025*
Pregnancy	6 (1%)	7 (1%)	1.10 (0.37–3.27)	.871
Smoker	30 (3%)	37 (3%)	1.16 (0.71–1.90)	.547
Chest X-ray infiltrates	626 (69%)	635 (66%)	0.87 (0.72–1.06)	.162
First antibiotic prescription before 24 h of admission
Any antibiotics	520 (49%)	304 (27%)	0.38 (0.32–0.46)	<.001
Narrow-spectrum^b	240 (23%)	154 (14%)	0.54 (0.43–0.67)	<.001*
Broad-spectrum^c	298 (28%)	157 (14%)	0.41 (0.33–0.51)	<.001*
First antibiotic prescription after 24 h of admission
Any antibiotics	163 (15%)	164 (15%)	0.93 (0.73–1.18)	.570
Narrow-spectrum^b	68(6%)	80 (7%)	1.11 (0.80–1.55)	.536
Broad-spectrum^c	114 (11%)	106 (9%)	0.86 (0.65–1.14)	.285
Total antibiotics
Any antibiotics	683 (65%)	468 (42%)	0.39 (0.33–0.46)	<.001*
Narrow-spectrum^b	344 (33%)	260 (23%)	0.62 (0.52–0.75)	<.001*
Broad-spectrum^c	490 (46%)	289 (26%)	0.86 (0.65–1.14)	<.001*

The first pandemic wave was defined as the time-period from March to June 2020, and the second pandemic wave as the time-period from July 2020 to January 2021.

^aMann–Whitney U test.

^bPenicillins without penicillinase-activity and aminoglycosides.

^cCarbapenems, cephalosporins, macrolides, tetracyclines, quinolones, piperacillin/tazobactam and others.

*p-value <.05. p-values < = .05 were considered significant. Chi-square statistics were used unless otherwise noted.

Table 3. Factors associated with antibiotic prescription.

	n	Antibiotic prescription (%)	OR (95%CI)	p-Value	aOR (95%CI)	p-Value
Diagnosis
COVID-19	82	40 (49%)	0.43 (0.26–0.73)	.002*	0.21 (0.09–0.50)	<.001*
Influenza	215	148 (69%)
DEMOGRAPHICS
Age^a	297	188 (63%)	1.03 (1.02–1.04)	.001*	1.01 (1.00–1.03)	.155*
Age groups
Older (≥ 65 years)	141	105 (75%)	2.57 (1.57–4.20)	<.001*
Younger(<65years)	156	83 (56%)
10–19	2	1 (50%)	1.58 (0.90–27.78)	.753
20–29	29	13 (45%)	1.29 (0.46–3.60)	.631
30–39 (ref)	31	12 (39%)	ref
40–49	28	12 (43%)	1.19 (0.42–3.36)	.746
50–49	43	29 (67%)	3.28 (1.25–8.60)	.016*
60–69	49	34 (69%)	3.59 (1.40–9.23)	.008*
70–79	54	43 (80%)	6.19 (2.32–16.50)	<.001*
80–89	46	31 (67%)	3.27 (1.27–8.46)	.014*
90–99	15	13 (87%)	10.29 (1.97–53.85)	.006*
Sex
Female	144	90 (62%)	0.94 (0.58–1.50)	.781	–
Male	153	98 (64%)
Comorbidities
Present	228	154 (68%)	2.21 (1.27–3.83)	.005*	1.16 (0.54–2.50)	.705
Absent	68	33 (49%)
Cardiovascular disease
Present	88	67 (76%)	2.32 (1.32–4.07)	.003*	–
Absent	209	121 (58%)
Hypertension
Present	94	69 (73%)	1.95 (1.14–3.33)	.015*	–
Absent	203	119 (59%)
Chronic lung disease
Present	104	72 (69%)	1.49 (0.90–2.48)	.120	–
Absent	193	116 (60%)
Smoking
Current	39	30 (77%)	2.11 (0.96–4.63)	.063	–
Previously or never	258	158 (61%)
Obesity (BMI > 30)
Present	52	32 (62%)	0.90 (0.48–1.69)	.747	–
Absent	197	126 (64%)
Diabetes Mellitus
Present	38	28 (74%)	1.73 (0.81–3.72)	.159	–
Absent	259	160 (62%)
Chronic renal disease
Present	33	24 (72%)	1.62 (0.73–3.64)	.237	–
Absent	264	164 (62%)
Chronic neurological disease
Present	57	40 (70%)	1.46 (0.78–2.73)	.233	–
Absent	240	148 (62%)
Immunosuppression
Present	32	26 (81%)	2.76 (1.10–6.92)	.031*
Absent	265	162 (61%)
Active cancer
Present	23	17 (74%)	1.70 (0.65–4.47)	.276	–
Absent	274	171 (64%)
Clinical presentation
Time from symptoms to admission^a
Days	293		0.91 (0.85–0.96)	.002*	0.93 (0.84–1.02)	.103
Temperature > 37.5 °C
Present	249	162 (65%)	1.58 (0.84–2.94)	.154	–
Absent	48	26 (54%)
Diagnostics
Chest X-ray
Infiltrate	120	94 (78%)	2.51 (1.45–4.36)	.001*	4.39 (1.94–9.93)	<.001*
No infiltrate	139	82 (59%)
Interventions
Oxygen treatment
Received	131	107 (82%)	4.74 (2.77–8.11)	<.001*	2.88 (1.49–5.57)	.002*
Not received	165	80 (49%)
NIV treatment
Received	36	36 (100%)	–	–	–
Not received	260	151 (58%)
Respirator treatment
Received	13	13 (100%)	–	–	–
Not received	282	174 (62%)

Antibiotic prescription was defined as the dependent variable. Independent variables entered in multiple logistic regression analysis were ‘diagnosis’, ‘age’, ‘comorbidities’, ‘duration of symptoms’, ‘chest X-ray infiltrates’ and ‘oxygen treatment’.

^aContinuous variables. Approximate percentage of variance accounted for in multivariable analysis was 25% (Cox & Snell R²=0.213 and Nagelkerke R²=0.299).