In our study, 1199 PCR positive fecal samples were included over the course of two years (Figure 2). From the fecal samples, 414 isolates were cultured and initially identified as 232 S. sonnei, 100 S. flexneri, 64 EIEC, 10 provisional Shigella, 3 S. boydii, for the remaining 5 isolates a distinction between S. flexneri and EIEC could not be made. Shigella were called provisional if the serotype could not be determined, or if the established serotype did not match with the phenotype. In total, 777 (65%) patients provided clinical and epidemiological data. Samples of these patients were included for the comparisons described below (Figure 2). In total, 290 of the 777 patients had a culture-positive infection. The data of patients from whom a S. sonnei, S. flexneri, S. boydii or provisional Shigella (n =255) isolate was obtained were used in the comparison to patients of whom an EIEC isolate (n=33) was cultured (Figure 2). For comparison of culture positive cases to culture negative cases, only data from patients of which S. sonnei or S. flexneri was cultured (n = 245) were compared to patients of which S. sonnei or S. flexneri was molecularly detected (n = 167) (Figure 2). One S. flexneri and one EIEC isolate were excluded from all analyses because they were cultured from the same fecal sample.
Assessment of the sensitivity and specificity of the molecular S. flexneri and S. sonnei serotyping directly from fecal samples resulted in a sensitivity of 77% and 75%, and a specificity of 98% and 99% respectively.
Incidence
During 2016 and 2017, 873 cases of shigellosis were notified to the health authorities, resulting in an average of 436.5 cases each year. The total number of residents in the Netherlands on 1 January 2017 was 17,081,507, resulting in an estimated incidence 135 shigellosis cases per 100,000 residents per year in the Netherlands during 2016 and 2017. Almost forty percent (39.5%) of all notified shigellosis cases were included in this study. We assumed the same ratio of EIEC cases having been included in our study and multiplied their number by 2.53 to estimate the national EIEC incidence rates. As 64 EIEC isolates were cultured, this resulted in 160 EIEC cases in 2 years, i.e., 80 per year. From the estimation for specific EIEC community incidence followed that a multiplier of 265 should be applied, see Additional File 1 for calculations. This resulted in 80*265 = 21,200 cases in the Dutch population, translated to a community incidence for EIEC of 124 cases per 100,000 residents per year in the Netherlands during 2016 and 2017.
Risks factors
Our results showed that patients with EIEC infections were more likely to report ingestion of suspected contaminated food or water (OR: 3.04 (1.44-6.42)) and less likely to report MSM contact (OR: 0.21 (0.05-0.98)) as source for infection compared to patients with Shigella spp. (Table 1).
As expected, Ct-values were approximately three Ct lower for the culture-positive shigellosis cases (OR: 0.88 (0.84-0.93)) than for culture-negative cases. Additionally, the proportion of S. flexneri in culture-positive infections was lower than the proportion in culture-negative infections (OR: 0.32 (0.19-0.54)). Furthermore, assessment of risk factors revealed that culture-positive cases travelled less (OR: 0.40 (0.20-0.78)) and were more likely to report MSM contact (OR: 3.22 (1.70-6.09)) or an unknown infection source (OR: 1.85 (1.17-2.92)) than culture-negative cases. In addition, culture-positive cases were less likely to report ingestion of suspected contaminated food or water as infection source than culture-negative cases (OR: 0.38 (0.24-0.61)) (Table 1).
Table 1 Risk factors of infections with EIEC and Shigella, and culture-positive and culture-negative shigellosis
Risk factors
|
EIECa, b (n=32)
|
Shigella spp.a
(n=254)
|
Univariate OR (95% CI)
|
Multivariate OR (95% CI)
|
Culture +/ PCR +a, b (n=244)
|
Culture - / PCR +
(n=167)
|
Univariate OR (95% CI)
|
Multivariate OR (95% CI)
|
Sex of patient (female)
|
44%
|
46%
|
0.91 (0.43-1.91)
|
|
46%
|
53%
|
0.76 (0.50-1.16)
|
|
Age of patient (mean ± sd)
|
36.0 ± 20.4
|
38.9 ± 18.5
|
0.99 (0.97-1.01)
|
|
38.7 ± 18.8
|
41.1 ± 19.3
|
0.99 (0.98-1.00)
|
|
Living in multi-person household
|
78%
|
74%
|
1.37 (0.57-3.33)
|
|
75%
|
80%
|
0.89 (0.58-1.35)
|
|
Co-infection with other enteric pathogen
|
28%
|
13%
|
2.72 (1.15-6.38)
|
|
12%
|
11%
|
1.04 (0.54-1.99)
|
|
Bacterial load (Ct-value, mean ± sd)
|
|
|
|
|
22.9 ± 4.6
|
25.3 ± 4.8
|
0.90 (0.86-0.94)
|
0.88 (0.84-0.93)
|
Species (S. flexneri)
|
|
|
|
|
31%
|
55%
|
0.36 (0.23-0.55)
|
0.32 (0.19-0.54)
|
Effect underlying disease/use of medication
|
|
|
|
|
- Higher infection risk
|
3%
|
20%
|
0.18 (0.03-0.91)
|
|
21%
|
17%
|
1.31 (0.82-2.08)
|
|
- More severe course
|
13%
|
7%
|
1.90 (0.69-5.20)
|
|
7%
|
6%
|
1.28 (0.65-2.55)
|
|
- Higher infection risk + more severe course
|
9%
|
10%
|
1.04 (0.35-3.07)
|
|
9%
|
11%
|
0.82 (0.46-1.46)
|
|
- Unknown effect
|
13%
|
6%
|
2.25 (0.81-6.24)
|
|
7%
|
11%
|
0.71 (0.39-1.31)
|
|
Travel history
|
88%
|
60%
|
4.62 (1.57-13.57)
|
|
57%
|
83%
|
0.26 (0.16-0.43)
|
0.40 (0.20-0.78)
|
Regions:
|
|
|
|
|
- South America
|
13%
|
4%
|
3.07 (1.04-9.04)
|
|
3%
|
5%
|
0.65 (0.25-1.69)
|
|
- Central America
|
13%
|
6%
|
1.73 (0.62-4.79)
|
|
5%
|
5%
|
0.95 (0.41-2.19)
|
|
- Asia
|
34%
|
17%
|
1.77 (0.85-3.67)
|
|
12%
|
26%
|
0.45 (0.26-0.77)
|
|
- Africa
|
25%
|
28%
|
0.79 (0.36-1.71)
|
|
30%
|
44%
|
0.65 (0.41-1.01)
|
|
- Europe
|
3%
|
6%
|
0.49 (0.09-2.78)
|
|
5%
|
2%
|
2.53 (0.84-7.68)
|
|
Source of infection (suspected by patient):
|
|
|
|
|
- Contaminated food/water
|
53%
|
26%
|
3.04 (1.44-6.42)
|
3.04 (1.44-6.42)
|
27%
|
64%
|
0.33 (0.23-0.48)
|
0.38 (0.24-0.61)
|
- MSM contact
|
3%
|
22%
|
0.21 (0.05-0.98)
|
0.21 (0.05-0.98)
|
24%
|
7%
|
2.84 (1.65-4.90)
|
3.22 (1.70-6.09)
|
- Unknown
|
38%
|
45%
|
1.25 (0.58-2.71)
|
1.25 (0.58-2.71)
|
42%
|
20%
|
1.70 (1.14-2.54)
|
1.85 (1.17-2.92)
|
Infection occupation related
|
9%
|
4%
|
1.64 (0.83-3.25)
|
|
3%
|
8%
|
0.62 (0.38-1.03)
|
|
OR = odds ratio, CI= 95% confidence interval, sd = standard deviation. aone S. flexneri and one EIEC isolate were excluded from analysis, because they caused a double-infection. bEIEC and culture + /PCR + were considered as cases, Shigella spp. and culture -/ PCR + as controls. Bold values indicate significant results with p-values < 0.05.
Symptoms, severity of disease and socio-economic consequences
Patients with EIEC infections reported suffering for longer from diarrhea than patients with Shigella spp. infection. In addition, the maximum vomiting frequency was higher for patients with EIEC infections (Table 2). Although patients with EIEC were symptomatic longer, they exhibited fewer symptoms and scoring lower on the de Wit scale than patients with Shigella spp. In contrast, no significant difference in severity was calculated using the MVS scale (Table 2). For socio-economic consequences, patients with EIEC infections were more likely to visit a general practitioner (GP) and to have a shorter stay when hospitalized than patients with a Shigella spp. infection (Table 3).
Table 2. Symptoms and severity of infections with EIEC and Shigella, and culture-positive and culture-negative shigellosis
Symptoms and severity
|
EIECa, b (n=32)
|
Shigella spp.a (n=254)
|
Univariate model,
p-value
|
Multivariate model,
p-value
|
Culture +/ PCR +a, b (n=244)
|
Culture - / PCR + (n=167)
|
Univariate model,
p-value
|
Multivariate model,
p-value
|
Blood in stool (% present)
|
16
|
39
|
0.005
|
0.051
|
39
|
38
|
0.901
|
0.679
|
Mucus in stool (% present)
|
47
|
58
|
0.222
|
0.290
|
58
|
54
|
0.508
|
0.688
|
Abdominal pain (% present)
|
59
|
74
|
0.082
|
0.108
|
75
|
71
|
0.330
|
0.945
|
Abdominal cramps (% present)
|
72
|
82
|
0.194
|
0.115
|
82
|
83
|
0.662
|
0.310
|
Nausea (% present)
|
56
|
44
|
0.209
|
0.568
|
45
|
54
|
0.066
|
0.041
|
Headache (% present)
|
22
|
33
|
0.187
|
0.052
|
32
|
40
|
0.108
|
0.086
|
Fever (% present)
|
47
|
60
|
0.164
|
0.248
|
59
|
56
|
0.582
|
0.420
|
When fever, duration in days
(median (IQR))
|
3 (2.5-4.5)
|
2 (1-4)
|
0.334
|
0.165
|
2 (1-4)
|
2 (1-4)
|
0.802
|
0.698
|
When fever, maximum temperature (mean ± sd)
|
40.0 ± 0.7
|
39.4 ± 0.9
|
0.063
|
0.413
|
39.4 ± 0.9
|
39.2 ± 0.8
|
0.084
|
0.179
|
Diarrhea (% present)
|
97
|
97
|
0.907
|
0.776
|
98
|
99
|
0.349
|
0.303
|
When diarrhea, duration in days (median (IQR))
|
14 (7-19.5)
|
10 (6-14)
|
<0.001
|
<0.001
|
9.5 (6-14)
|
14 (8-24)
|
<0.001
|
0.001
|
When diarrhea, frequency in 24H (median (IQR))
|
8 (6-14)
|
9 (6-15)
|
0.855
|
0.796
|
10 (6-15)
|
10 (6-16)
|
0.486
|
0.185
|
Vomiting (% present)
|
28
|
28
|
0.979
|
0.809
|
29
|
37
|
0.073
|
0.026
|
When vomiting, duration in days
(median (IQR))
|
2 (1-3)
|
1 (1-3)
|
0.508
|
0.929
|
1 (1-3)
|
2 (1-3)
|
0.033
|
0.167
|
When vomiting, frequency in 24H
(median (IQR))
|
3 (2-8)
|
2 (1-4)
|
0.166
|
0.001
|
2 (1-4)
|
3 (1-5.8)
|
0.525
|
0.027
|
Total number of symptoms (median (IQR))
|
4 (3.0-5.3)
|
5 (4-6)
|
0.006
|
0.006
|
5 (4-6)
|
5 (4-6)
|
0.519
|
0.104
|
Severity scores:
|
|
|
|
|
|
|
|
|
- de Wit et al.
(mean ± sd)
|
6.4 ±2.6
|
7.5 ±2.7
|
0.033
|
0.045
|
7.5 ± 2.7
|
7.7 ± 2.7
|
0.380
|
0.132
|
- Modified vesikari (mean ± sd)
|
7.4 ± 3.3
|
7.3 ± 2.8
|
0.852
|
0.943
|
7.3 ± 2.8
|
7.9 ± 2.8
|
0.028
|
0.007
|
Sd = standard deviation, IQR = interquartile range. aone S. flexneri and one EIEC isolate were excluded from analysis, because they caused a double-infection. Bold values indicate significant results with p-values < 0.05.
Culture-negative cases were more likely to report nausea, longer duration of diarrhea, vomiting and higher frequencies of vomiting than culture-positive cases (Table 2). Moreover, the MVS score of culture-negative cases was significantly higher than that of culture-positive cases, while the de Wit scores showed no significant difference (Table 2). In addition, culture-negative cases were more likely to report longer absence from work compared to culture-positive cases (Table 3).
Table 3. Socio-economic consequences of infections with EIEC and Shigella, and culture-positive and culture-negative shigellosis
Consequences
|
EIECa, b (n=32)
|
Shigella spp.a (n=254)
|
Univariate model, p-value
|
Multivariate model, p-value
|
Culture +/ PCR +a, b (n=244)
|
Culture - / PCR + (n=167)
|
Univariate model, p-value
|
Multivariate model, p-value
|
Bedrest (% present)
|
88
|
81
|
0.357
|
0.186
|
82
|
79
|
0.528
|
0.514
|
Leave of absence (% present)
|
56
|
53
|
0.709
|
0.703
|
53
|
47
|
0.220
|
0.737
|
When absence patient, duration in days (median (IQR))
|
5 (3.0-7.8)
|
4 (3-7)
|
0.882
|
0.401
|
4 (3-7)
|
7 (3-10)
|
0.038
|
0.005
|
When absence caretaker,
duration in days (median (IQR))
|
0 (0-0)
|
0 (0-0)
|
0.554
|
0.185
|
0 (0-0)
|
0 (0-0)
|
0.389
|
0.171
|
Use of care facilities
|
|
|
|
|
GP (% visited)
|
100
|
91
|
0.015
|
0.037
|
91
|
93
|
0.299
|
0.851
|
When GP visited, number of visits
(median (IQR))
|
1.5 (1-2)
|
1 (1-2)
|
0.623
|
0.399
|
1 (1-2)
|
1 (1-2)
|
0.595
|
0.909
|
GP outside office hours (% visited)
|
9
|
9
|
0.989
|
0.537
|
9
|
10
|
0.694
|
0.757
|
Specialists (% visited)
|
16
|
13
|
0.732
|
0.830
|
13
|
16
|
0.388
|
0.965
|
When specialist visited, number of visits (median (IQR))
|
1 (1-2)
|
1 (1-1)
|
0.797
|
0.799
|
1 (1-1)
|
1 (1-2)
|
0.122
|
0.553
|
Emergency room (% visited)
|
9
|
10
|
0.933
|
0.781
|
10
|
5
|
0.072
|
0.074
|
Hospitalization (% hospitalized)
|
3
|
9
|
0.180
|
0.270
|
9
|
5
|
0.163
|
0.443
|
When hospitalized, duration in
days (median (IQR))
|
1.5 (0.8-2.3)
|
3 (2-4)
|
0.179
|
0.027
|
3 (1.5-3.5)
|
3.5 (1-4.8)
|
0.244
|
0.648
|
IQR = interquartile range. aone S. flexneri and one EIEC isolate were excluded from analysis, because they caused a double-infection. Bold values indicate significant results with p-values < 0.05.
Secondary infections
Because there was a lack of specific data about relationships between cases, the presence and number of self-reported related patients was used as a proxy for the degree of secondary infections. No significant differences in presence and number of self-reported related patients were found when comparing EIEC cases with shigellosis cases or when comparing culture-positive cases to culture-negative cases (Table 4).
Table 4. Degree secondary infections of infections with EIEC and Shigella, and culture-positive and culture-negative shigellosis
Secondary infections
|
EIECa, b (n=32)
|
Shigella spp.a (n=254)
|
Univariate model, p-value
|
Multivariate model, p-value
|
Culture +/ PCR +a, b (n=244)
|
Culture - / PCR + (n=167)
|
Univariate model, p-value
|
Multivariate model, p-value
|
Related patients (% present)
|
47
|
39
|
0.393
|
0.785
|
40
|
39
|
0.865
|
0.930
|
When related patients, total number (median (IQR))
|
1 (1-2)
|
1 (1-2)
|
0.239
|
0.354
|
1 (1-2)
|
1 (1-3)
|
0.326
|
0.977
|
IQR = interquartile range. aone S. flexneri and one EIEC isolate were excluded from analysis, because they caused a double-infection. Bold values indicate significant results with p-values < 0.05
Comparison of infections with cultured S. flexneri and S. sonnei
First, patients with S. sonnei were more likely to report (85%) abdominal cramps compared to S. flexneri (75%, p = 0.047). Second, no differences in total number of symptoms or disease severity were found. Third, patients with S. sonnei were more likely to self-report the presence of related patients (45%) than patients with S. flexneri (28%, p = 0.028), although the self-reported number of related patients did not differ. Fourth, for the socio-economic consequences, there were multiple differences: patients with S. flexneri were more likely to report longer absence from work (median 5 (3-9) days), multiple visits to their GP (average 2.1 visits), visits to specialists (21%) and hospitalization (17%) compared to patients with S. sonnei that reported a median of 4 (2-7), p = 0.001)) days of absence, an average of 1.6 GP visits (p = 0.049), 10% specialist visits (p = 0.015), and 5% hospitalization (p < 0.001).