Journal
of Hospital
Infection
A review
(1992)
20,
Tyneside
11
of hospitalized patients gastroenteritis G. Gopal
North
105-l
General
Hospital,
Rao and
bacterial
M. Fuller
Rake Lane, NE29 8NH
Accepted for publication
with
North Shields, Tyne and Wear
5 December 1991
Summary:
The distribution and clinical management of thirty-two hospitalized patients with salmonella and campylobacter infections were reviewed and the impact of these infections on hospital resources was assessed. Eighteen patients with salmonella infection had an age and sex distribution comparable with the community cases. In contrast, 10 out of 14 (71.4%) patients with campylobacter infection were under 20 years of age though the peak incidence of the infection in the community occurred in the 21-65 years age group (67%). There was no male predominance. The median duration of stay in hospital was 6 days for patients with salmonella infection and 3 days for those with campylobacter infection. Physicians were inconsistent in the treatment of campylobacter infection. Overall the financial impact of managing patients with salmonella and campylobacter infection was considerable (El384 and A779 respectively per patient). A limitation on unnecessarily prolonged hospital stays and the establishment of clear guidelines for the clinical management of these infections are necessary. Keywords:
Campylobacter;
Salmonella;
hospitalized
patients.
Introduction
Campylobacter spp. and Salmonella spp. are the commonest causes of bacterial gastroenteritis in the UK. In 1989, 32 000 campylobacter and 29 998 salmonella infections were reported from England and Wales to the Central Public Health Laboratory, London.’ This was higher than in any of the previous years. Our microbiology laboratory is attached to a 700-bed district general hospital and serves both the hospital and the general practitioners in the district. During 1989 an increase was observed in the number of isolations of Campylobacter and Salmonella compared with previous years and, during the same period, more cases of campylobacter and salmonella infection were admitted to the hospital. We therefore reviewed patients with campylobacter and salmonella infections who were admitted to the hospital in order to compare the clinical management of these patients and to assess the impact of these two infections on hospital Correspondence
to: Dr
G. Gopal
019%6701/92/020105
+07 $03 00/O
Rao 0 1992 The Hospital
105
Infection
Society
106
G. Gopal
Rao and M. Fuller
resources. It was hoped that such a review might identify differences between the two infections which would have helped in the management of these patients. Another aim was to highlight areas where clinical practice could be altered in order to make more efficient and rational use of hospital resources. Methods
Case records of patients admitted to the hospital in 1989 with campylobacter or salmonella infections were reviewed. The following information was gathered; age, sex, duration of stay, number of days in isolation, investigations and procedures performed, treatment, outcome and the organism isolated. An attempt was made to exclude investigations or procedures performed for co-existing conditions, unrelated to campylobacter or salmonella infections. Costs of hospitalization, investigations and procedures were obtained from the hospital finance department. Results
In 1989, 142 cases of campylobacter and 105 of salmonella infection were identified in the laboratory. Of these, 36 (14.5%) patients were admitted to the hospital. Case records of 14 of the 15 patients with campylobacter infections and 18 of the 21 patients with salmonella infections were available for analysis. The age distribution of all the infected patients is given in Figure 1. Figure 2 shows the age distribution of the hospitalized patients. Ten of the 14 (71.4%) campylobacter patients admitted to hospital were under 20 years of age although that age group represented only 19.4% of all the infections recorded in the laboratory (Figure 2). The age distribution of all our patients with Campylobacter suggested that the infection was most common in the 21-40-year age group (47%) followed by the 41-65-year age group (20%). Compared with campylobacter infections, a higher proportion of patients with salmonella infection were admitted to the hospital (17%). Three other patients developed the infection whilst in hospital. Table I compares the median duration of stay, clinical presentation, investigation, procedures, antimicrobial treatment and costs in the two types of infection. Although diarrhoea (64% and 68%) and abdominal pain (27% and 28%) were the commonest presentations of both campylobacter and salmonella infections, there were several important differences between the two. Presence of blood in the stool or frank bleeding per rectum was more common in campylobacter infection (35.7%) than that of Salmonella (11.1%). This probably reflects the greater propensity of Campylobacter to cause enterocolitis. Another notable feature in the presentation of
Ir
Bacterial
gastroenteritis
in hospital
107
7(
6(
6-10
I I-20 Age
Figure
1. Age
distribution
1989. E8 = Campylobactor,
of campylobacter
distrlbutlon
and
41-65
765
(years)
salmonella
infections
in the community
in
q = Salmonella.
campylobacter infection was that two of the 14 cases admitted presented as ‘pseudoappendicitis’. Patients with campylobacter infection had a median duration of stay of 3 days (mean of 6.5 days) compared with 6 days (mean of 11.9 days) for salmonella infections. In the majority of cases of salmonella gastroenteritis, the clinicians withheld antimicrobial treatment. The only occasion where specific salmonella treatment was given was to a child with salmonella arthritis. In contrast, there were marked differences in the frequency of antimicrobial treatment of campylobacter infections in adults and children. Whilst only two of six (33.3%) adults received erythromycin, all eight children were treated.
G. Gopal
l-5
II-XI
b-IU Age
Figure 1989.
2. Age distribution Ed = Campylobacter,
Rao and M. Fuller
of campylobacter W = Salmonella.
distribution
and
41-64
>65
(years)
salmonella
infections
in the
hospital
in
Ail patients admitted with campylobacter infection had fully recovered or were recovering at the time of discharge. Three of the 18 patients with salmonella infection died during hospitalization. Two of the three had underlying diseases which were the likely cause of death. It was, however, not possible to establish whether SaZmoneZZacontributed to the death of the third person. One pregnant patient who was admitted at term with campylobacter infection had a normal delivery. Overall, there were more patients with salmonella infections admitted to hospital, a higher proportion of them were elderly, more investigations were performed and their overall median duration of stay was almost twice as
Bacterial Table
I. Details
relating
gastroenteritis
to hospitalized
patients batter
Feature
in hospital infected spp.
with
Salmonella
Median
duration
of stay
in days
Number of patients presenting Diarrhoea and vomiting Abdominal pain Blood in stools and bleeding Pseudoappendicitis Others
(range)
6 (3-36)
per
or Campylo-
Campylobacter 3 (l-12)
per
rectum
12 5 2 0 6 72 42 38 11 7 1
27 9 4 5 9 2
1 0
0 2
Number of procedures Colonoscopy Appendicectomy Number
Salmonella
with
Number of investigations Microbiology Biochemistry Haematology Radiology Virology Histology
Cost
either
109
patients
given
antimicrobial
patient
(pounds
sterling)
treatment
1 1384
11 779
long as that of patients with campylobacter infections. The duration of stay in the hospital was the most significant factor which added to the cost of treatment. The overall cost was estimated to be E3.5 821. The cost of treating salmonella infections was significantly greater than that for campylobacter infections (;1;1384 compared with A779 per patient). Discussion
In this study we reviewed the majority of the cases of campylobacter and salmonella infections admitted to our hospital in a year when these infections occurred at a record high rate locally. The fact that 15 out of 142 (10.5%) campylobacter and 18 out of 105 (17%) salmonella infections identified in the laboratory resulted in hospital admission gives an indication of the potential severity of these infections. This is particularly important in a population where the proportion of elderly (> 6.5 years) and very young people (< 5 years) is high because of the potential severity of such infection in these groups. Although a bimodal distribution of campylobacter infection was not seen, this study confirms that the highest incidence of the infection is in young adults.’ This disparity between age of maximum incidence of campylobacter infection and the age of those admitted to hospital may be because infection is particularly severe in the very young, and adults may have a relatively less severe illness or be able to cope with their illness at home. The age distribution of patients admitted to
110
G. Gopal
Rao and M. Fuller
hospital with salmonella infections was broadly similar to that of all patients identified in the laboratory. There was no significant difference in the incidence of campylobacter and salmonella infections between the sexes. The increased incidence of campylobacter infection in males3 was not seen in this study probably because of the small patient numbers. The presentation of campylobacter infection suggesting appendicitis has been described before.4 Although campylobacter infection should be considered in the differential diagnosis of young patients presenting signs of appendicitis, it is difficult to suggest how this could be confirmed before the onset of diarrhoea. Once diarrhoea occurs, rapid dark-ground microscopy of the stool may enable a presumptive diagnosis. However, few laboratories in the UK perform direct microscopy for detecting campylobacter infection. One child with salmonella infection presented with septic arthritis of the hip. Although such a presentation is uncommon in patients without predisposing factors such as sickle cell disease, clinicians must be aware that salmonella arthritis can occur in otherwise healthy children. The reason for the difference between the lengths of stay in hospital of the two groups of patients is unclear. A variety of factors could be responsible including the routine practices of the attending physicians, the age and the clinical condition of the patients. The majority of children with campylobacter infections (eight out of 14) had a mean stay in hospital of 2.42 days in contrast to the older patients whose mean stay was 5.7 days. This difference may reflect either the greater familiarity of the paediatricians with gastroenteritis or the need to consider a wider differential diagnosis in adults presenting with gastrointestinal symptoms. The latter approach wou,ld result in the greater frequency and range of investigations recorded in older patients. A disquieting feature which emerged from the analysis, was that some clinicians insisted on obtaining negative stool cultures, even from patients with formed stools, before discharging them home. This is usually unnecessary, because they are unlikely to transmit infection unless they fall into one of the high-risk categories, such as food handlers, children attending nurseries and immunosuppressed patients. The extended hospitalization of otherwise asymptomatic patients in order to obtain negative stool culture seems to be a waste of resources. In this context, it is of interest that although antibiotics are generally not recommended in the management of uncomplicated salmonella infection in immunocompetent patients, quinolones have been used successfully in the control of outbreaks of salmonella infection in hospital.5 However, reports of the emergence of resistance in bacteria to quinolones emphasizes the need to use these agents with caution6 Treatment for campylobacter infection is indicated when the patient has a high fever, bloody diarrhoea or passes more than eight stools per day. It is
Bacterial
gastroenteritis
in
hospital
111
also necessary in patients whose symptoms are unimproved or worsening at the time the diagnosis is made or in those whose symptoms have persisted for more than 1 week.7 Considering that only four of the eight children and three of the six adults fulfilled the above criteria, the basis for initiation of treatment is unclear. One of the reasons for treating all children could be that some studies on children have shown a clear benefit from early treatment. We were unable to assess the outcome of treatment because the follow-up of the patients was left to the general practitioners and records were not available for analysis. From the foregoing, it would appear that although various studies have suggested indications for treatment, they are not adhered to uniformly by clinicians. In our attempt to assess the economic impact of campylobacter and salmonella infections on the hospital resources, we are aware of many limitations, such as the inability to estimate accurately the cost of nursing patients in isolation. We used only broad areas of assessrnent of the costs. These included the basic cost of occupying a hospital bed, and the cost of investigations and procedures. One of the consequences of the overall increase in campylobacter and salmonella infections in the community is that more patients are admitted to hospital, with important implications for hospital resources. It is therefore important to look closely at the management of these infections in hospital. There appears to be a wide disparity in the way physicians perceive these infections and we believe there is a need to establish clear guidelines for the clinical management of these patients. Particular attention needs to be paid to unnecessarily prolonged hospitalization. The authors preparation
would like of this paper.
to
thank
Mrs
C.
M.
Chicken
for
secretarial
assistance
in
the
References 1. Cooke EM. Epidemiology of foodborne illness. Lancet 1990; 336: 79c-793. 2. Skirrow MB. Foodborne illness: Campylobacter. Lancet 1990; 336: 921-923. 3. Skirrow MB. A demographic study of campylobacter, salmonella and shigella infection in the UK. A Public Health Laboratory Study. Epidemiol Infect 1987; 99: 6477657. 4. Blasen MJ, Benkoenits ID, La Force M. Campylobacter enteritis; clinical and epidemiological features. Ann Intern Med 1979; 91: 179. 5. Ahmad F, Bray G, Prescott RWG, Aquilla S, Lightfoot NF. Use of ciprofloxacin to control a salmonella outbreak in a long-stay psychiatric hospital. J Hasp Infect 1991; 17: 171-178. 6. Howard AJ, Joseph TD, Bloodsworth LLO, Frost JA, Chart H, Rowe B. The emergence of ciprofloxacin resistance in Salmonella typhimurium. J Antimicrob Chemother 1990; 26: 296-298. 7. Salazan-Lindo E, Sack RB, Chea-Woo E et al. Early treatment with erythromycin of Campylobacter jejuni associated dysentery in children. J Pediatr 1986; 109: 355.
of Hospital
Infection
A review
(1992)
20,
Tyneside
11
of hospitalized patients gastroenteritis G. Gopal
North
105-l
General
Hospital,
Rao and
bacterial
M. Fuller
Rake Lane, NE29 8NH
Accepted for publication
with
North Shields, Tyne and Wear
5 December 1991
Summary:
The distribution and clinical management of thirty-two hospitalized patients with salmonella and campylobacter infections were reviewed and the impact of these infections on hospital resources was assessed. Eighteen patients with salmonella infection had an age and sex distribution comparable with the community cases. In contrast, 10 out of 14 (71.4%) patients with campylobacter infection were under 20 years of age though the peak incidence of the infection in the community occurred in the 21-65 years age group (67%). There was no male predominance. The median duration of stay in hospital was 6 days for patients with salmonella infection and 3 days for those with campylobacter infection. Physicians were inconsistent in the treatment of campylobacter infection. Overall the financial impact of managing patients with salmonella and campylobacter infection was considerable (El384 and A779 respectively per patient). A limitation on unnecessarily prolonged hospital stays and the establishment of clear guidelines for the clinical management of these infections are necessary. Keywords:
Campylobacter;
Salmonella;
hospitalized
patients.
Introduction
Campylobacter spp. and Salmonella spp. are the commonest causes of bacterial gastroenteritis in the UK. In 1989, 32 000 campylobacter and 29 998 salmonella infections were reported from England and Wales to the Central Public Health Laboratory, London.’ This was higher than in any of the previous years. Our microbiology laboratory is attached to a 700-bed district general hospital and serves both the hospital and the general practitioners in the district. During 1989 an increase was observed in the number of isolations of Campylobacter and Salmonella compared with previous years and, during the same period, more cases of campylobacter and salmonella infection were admitted to the hospital. We therefore reviewed patients with campylobacter and salmonella infections who were admitted to the hospital in order to compare the clinical management of these patients and to assess the impact of these two infections on hospital Correspondence
to: Dr
G. Gopal
019%6701/92/020105
+07 $03 00/O
Rao 0 1992 The Hospital
105
Infection
Society
106
G. Gopal
Rao and M. Fuller
resources. It was hoped that such a review might identify differences between the two infections which would have helped in the management of these patients. Another aim was to highlight areas where clinical practice could be altered in order to make more efficient and rational use of hospital resources. Methods
Case records of patients admitted to the hospital in 1989 with campylobacter or salmonella infections were reviewed. The following information was gathered; age, sex, duration of stay, number of days in isolation, investigations and procedures performed, treatment, outcome and the organism isolated. An attempt was made to exclude investigations or procedures performed for co-existing conditions, unrelated to campylobacter or salmonella infections. Costs of hospitalization, investigations and procedures were obtained from the hospital finance department. Results
In 1989, 142 cases of campylobacter and 105 of salmonella infection were identified in the laboratory. Of these, 36 (14.5%) patients were admitted to the hospital. Case records of 14 of the 15 patients with campylobacter infections and 18 of the 21 patients with salmonella infections were available for analysis. The age distribution of all the infected patients is given in Figure 1. Figure 2 shows the age distribution of the hospitalized patients. Ten of the 14 (71.4%) campylobacter patients admitted to hospital were under 20 years of age although that age group represented only 19.4% of all the infections recorded in the laboratory (Figure 2). The age distribution of all our patients with Campylobacter suggested that the infection was most common in the 21-40-year age group (47%) followed by the 41-65-year age group (20%). Compared with campylobacter infections, a higher proportion of patients with salmonella infection were admitted to the hospital (17%). Three other patients developed the infection whilst in hospital. Table I compares the median duration of stay, clinical presentation, investigation, procedures, antimicrobial treatment and costs in the two types of infection. Although diarrhoea (64% and 68%) and abdominal pain (27% and 28%) were the commonest presentations of both campylobacter and salmonella infections, there were several important differences between the two. Presence of blood in the stool or frank bleeding per rectum was more common in campylobacter infection (35.7%) than that of Salmonella (11.1%). This probably reflects the greater propensity of Campylobacter to cause enterocolitis. Another notable feature in the presentation of
Ir
Bacterial
gastroenteritis
in hospital
107
7(
6(
6-10
I I-20 Age
Figure
1. Age
distribution
1989. E8 = Campylobactor,
of campylobacter
distrlbutlon
and
41-65
765
(years)
salmonella
infections
in the community
in
q = Salmonella.
campylobacter infection was that two of the 14 cases admitted presented as ‘pseudoappendicitis’. Patients with campylobacter infection had a median duration of stay of 3 days (mean of 6.5 days) compared with 6 days (mean of 11.9 days) for salmonella infections. In the majority of cases of salmonella gastroenteritis, the clinicians withheld antimicrobial treatment. The only occasion where specific salmonella treatment was given was to a child with salmonella arthritis. In contrast, there were marked differences in the frequency of antimicrobial treatment of campylobacter infections in adults and children. Whilst only two of six (33.3%) adults received erythromycin, all eight children were treated.
G. Gopal
l-5
II-XI
b-IU Age
Figure 1989.
2. Age distribution Ed = Campylobacter,
Rao and M. Fuller
of campylobacter W = Salmonella.
distribution
and
41-64
>65
(years)
salmonella
infections
in the
hospital
in
Ail patients admitted with campylobacter infection had fully recovered or were recovering at the time of discharge. Three of the 18 patients with salmonella infection died during hospitalization. Two of the three had underlying diseases which were the likely cause of death. It was, however, not possible to establish whether SaZmoneZZacontributed to the death of the third person. One pregnant patient who was admitted at term with campylobacter infection had a normal delivery. Overall, there were more patients with salmonella infections admitted to hospital, a higher proportion of them were elderly, more investigations were performed and their overall median duration of stay was almost twice as
Bacterial Table
I. Details
relating
gastroenteritis
to hospitalized
patients batter
Feature
in hospital infected spp.
with
Salmonella
Median
duration
of stay
in days
Number of patients presenting Diarrhoea and vomiting Abdominal pain Blood in stools and bleeding Pseudoappendicitis Others
(range)
6 (3-36)
per
or Campylo-
Campylobacter 3 (l-12)
per
rectum
12 5 2 0 6 72 42 38 11 7 1
27 9 4 5 9 2
1 0
0 2
Number of procedures Colonoscopy Appendicectomy Number
Salmonella
with
Number of investigations Microbiology Biochemistry Haematology Radiology Virology Histology
Cost
either
109
patients
given
antimicrobial
patient
(pounds
sterling)
treatment
1 1384
11 779
long as that of patients with campylobacter infections. The duration of stay in the hospital was the most significant factor which added to the cost of treatment. The overall cost was estimated to be E3.5 821. The cost of treating salmonella infections was significantly greater than that for campylobacter infections (;1;1384 compared with A779 per patient). Discussion
In this study we reviewed the majority of the cases of campylobacter and salmonella infections admitted to our hospital in a year when these infections occurred at a record high rate locally. The fact that 15 out of 142 (10.5%) campylobacter and 18 out of 105 (17%) salmonella infections identified in the laboratory resulted in hospital admission gives an indication of the potential severity of these infections. This is particularly important in a population where the proportion of elderly (> 6.5 years) and very young people (< 5 years) is high because of the potential severity of such infection in these groups. Although a bimodal distribution of campylobacter infection was not seen, this study confirms that the highest incidence of the infection is in young adults.’ This disparity between age of maximum incidence of campylobacter infection and the age of those admitted to hospital may be because infection is particularly severe in the very young, and adults may have a relatively less severe illness or be able to cope with their illness at home. The age distribution of patients admitted to
110
G. Gopal
Rao and M. Fuller
hospital with salmonella infections was broadly similar to that of all patients identified in the laboratory. There was no significant difference in the incidence of campylobacter and salmonella infections between the sexes. The increased incidence of campylobacter infection in males3 was not seen in this study probably because of the small patient numbers. The presentation of campylobacter infection suggesting appendicitis has been described before.4 Although campylobacter infection should be considered in the differential diagnosis of young patients presenting signs of appendicitis, it is difficult to suggest how this could be confirmed before the onset of diarrhoea. Once diarrhoea occurs, rapid dark-ground microscopy of the stool may enable a presumptive diagnosis. However, few laboratories in the UK perform direct microscopy for detecting campylobacter infection. One child with salmonella infection presented with septic arthritis of the hip. Although such a presentation is uncommon in patients without predisposing factors such as sickle cell disease, clinicians must be aware that salmonella arthritis can occur in otherwise healthy children. The reason for the difference between the lengths of stay in hospital of the two groups of patients is unclear. A variety of factors could be responsible including the routine practices of the attending physicians, the age and the clinical condition of the patients. The majority of children with campylobacter infections (eight out of 14) had a mean stay in hospital of 2.42 days in contrast to the older patients whose mean stay was 5.7 days. This difference may reflect either the greater familiarity of the paediatricians with gastroenteritis or the need to consider a wider differential diagnosis in adults presenting with gastrointestinal symptoms. The latter approach wou,ld result in the greater frequency and range of investigations recorded in older patients. A disquieting feature which emerged from the analysis, was that some clinicians insisted on obtaining negative stool cultures, even from patients with formed stools, before discharging them home. This is usually unnecessary, because they are unlikely to transmit infection unless they fall into one of the high-risk categories, such as food handlers, children attending nurseries and immunosuppressed patients. The extended hospitalization of otherwise asymptomatic patients in order to obtain negative stool culture seems to be a waste of resources. In this context, it is of interest that although antibiotics are generally not recommended in the management of uncomplicated salmonella infection in immunocompetent patients, quinolones have been used successfully in the control of outbreaks of salmonella infection in hospital.5 However, reports of the emergence of resistance in bacteria to quinolones emphasizes the need to use these agents with caution6 Treatment for campylobacter infection is indicated when the patient has a high fever, bloody diarrhoea or passes more than eight stools per day. It is
Bacterial
gastroenteritis
in
hospital
111
also necessary in patients whose symptoms are unimproved or worsening at the time the diagnosis is made or in those whose symptoms have persisted for more than 1 week.7 Considering that only four of the eight children and three of the six adults fulfilled the above criteria, the basis for initiation of treatment is unclear. One of the reasons for treating all children could be that some studies on children have shown a clear benefit from early treatment. We were unable to assess the outcome of treatment because the follow-up of the patients was left to the general practitioners and records were not available for analysis. From the foregoing, it would appear that although various studies have suggested indications for treatment, they are not adhered to uniformly by clinicians. In our attempt to assess the economic impact of campylobacter and salmonella infections on the hospital resources, we are aware of many limitations, such as the inability to estimate accurately the cost of nursing patients in isolation. We used only broad areas of assessrnent of the costs. These included the basic cost of occupying a hospital bed, and the cost of investigations and procedures. One of the consequences of the overall increase in campylobacter and salmonella infections in the community is that more patients are admitted to hospital, with important implications for hospital resources. It is therefore important to look closely at the management of these infections in hospital. There appears to be a wide disparity in the way physicians perceive these infections and we believe there is a need to establish clear guidelines for the clinical management of these patients. Particular attention needs to be paid to unnecessarily prolonged hospitalization. The authors preparation
would like of this paper.
to
thank
Mrs
C.
M.
Chicken
for
secretarial
assistance
in
the
References 1. Cooke EM. Epidemiology of foodborne illness. Lancet 1990; 336: 79c-793. 2. Skirrow MB. Foodborne illness: Campylobacter. Lancet 1990; 336: 921-923. 3. Skirrow MB. A demographic study of campylobacter, salmonella and shigella infection in the UK. A Public Health Laboratory Study. Epidemiol Infect 1987; 99: 6477657. 4. Blasen MJ, Benkoenits ID, La Force M. Campylobacter enteritis; clinical and epidemiological features. Ann Intern Med 1979; 91: 179. 5. Ahmad F, Bray G, Prescott RWG, Aquilla S, Lightfoot NF. Use of ciprofloxacin to control a salmonella outbreak in a long-stay psychiatric hospital. J Hasp Infect 1991; 17: 171-178. 6. Howard AJ, Joseph TD, Bloodsworth LLO, Frost JA, Chart H, Rowe B. The emergence of ciprofloxacin resistance in Salmonella typhimurium. J Antimicrob Chemother 1990; 26: 296-298. 7. Salazan-Lindo E, Sack RB, Chea-Woo E et al. Early treatment with erythromycin of Campylobacter jejuni associated dysentery in children. J Pediatr 1986; 109: 355.
