Changing carriage rate of Neisseria meningitidis among university students during the first week of term: cross sectional study
Keith R Neal, Jonathan S Nguyen-Van-Tam, Nicholas
Jeffrey, Richard C B Slack, Richard J Madeley,
Kamel Ait-Tahar, Katy Job, Martin C J Wale,
Dlawer A A Ala'Aldeen
Abstract
Objective
To determine the rates of, and risk factors for, meningococcal carriage
and acquisition among university students.
Design
Repeated cross sectional study.
Participants
2507 students in their first year at university.
Main outcome measures
Prevalence of carriage of meningococci and risk factors for carriage
and acquisition of meningococci.
Results
Carriage rates for meningoccoci increased rapidly in the first week
of term from 6.9% on day 1, to 11.2% on day 2, to 19.0% on day 3,
and to 23.1% on day 4. The average carriage rate during the first
week of term in October among students living in catered halls was
13.9%. By November this had risen to 31.0% and in December it had
reached 34.2%. Independent associations for acquisition of meningococci
in the autumn term were frequency of visits to a hall bar (5.7 visits:
odds ratio 2.7, 95% confidence interval 1.5 to 4.8), active smoking
(1.6, 1.0 to 2.6), being male (1.6, 1.2 to 2.2), visits to night
clubs (1.3, 1.0 to 1.6), and intimate kissing (1.4, 1.0 to 1.8).
Lower rates of acquisition were found in female only halls (0.5,
0.3 to 0.9). The most commonly acquired meningococcal strain was
C2a P1.5 (P1.2), which has been implicated in clusters of invasive
meningococcal disease at other UK universities.
Conclusions
Carriage rates of meningococci among university students increase
rapidly in the first week of term, with further increases during
the term. The rapid rate of acquisition may explain the increased
risk of invasive meningococcal disease and the timing of cases and
outbreaks in university students.
Introduction
During the 1990s there have been major increases in the incidence
of invasive meningococcal disease in many developed countries, 1.3
with serogroup C disease most noticeable, especially among teenagers
and young adults. It has also been shown that university undergraduates
have higher rates of invasive meningococcal disease than young adults
of the same age who are not attending university.4 The
provision of places in catered halls seems to be an important factor
in differences in rates of invasive meningococcal disease between
universities.4 In the United Kingdom, several clusters
of invasive meningococcal disease have been reported; a large outbreak
occurred in November 1996 at the University of Wales in Cardiff5
and another in October 1997 at the University of Southampton.6
No studies have been published on the epidemiology of meningococcal
carriage or acquisition among university students in situations
where there are no outbreaks.7 We therefore performed
a longitudinal study in first year university students to determine
rates of carriage and acquisition of Neisseria meningitidis, together
with risk factors for both.
Participants and methods
Recruitment of students
Nottingham University is a large campus based institution. As part
of routine induction, all new students (mainly first year undergraduates)
are asked to attend the health centre on campus during their first
week at university. The order of attendance, evenly distributed
across the four days, is set by degree course and not by faculty
or hall of residence. During this week in October 1997, we recruited
students to the study after they had registered with the health
centre and undergone health screening. Each student was given an
information sheet and consent form. Those agreeing to take part
completed a questionnaire covering: personal characteristics, place
of residence, faculty, recent symptoms of upper respiratory tract
infection, medical history including meningococcal vaccination,
current and recent drugs, travel abroad and to other universities
in the past month; active and passive smoking, visits to bars and
night clubs, amount and type of alcohol consumed, number of people
kissed, and the sharing of glasses and cigarettes in the preceding
week. After each student had completed the questionnaire, a trained
operator took a posterior pharyngeal swab, which was plated immediately
on to selective medium and handled using standard tech. niques (see
website). The same processing methods were used throughout.
Follow up
All participants in catered halls were selected for a further pharyngeal
swab in either the first week of November 1997 or the first week
of December 1997 on the basis of odd or even study numbers. Pharyngeal
swabs were taken from students in the only self catered hall in
the study in December. At the time of reswabbing the questionnaire
was repeated. We therefore had paired data available for these students
for October and either November or December.
One case of serogroup C disease occurred in a catered hall in late
October, and one third of the students in this hall were therefore
given ciprofloxacin to eradicate meningococcal carriage. Although
stu. dents in this hall were reswabbed according to the pre-arranged
schedule, we excluded them from the main analysis. Two other cases
of invasive meningococcal disease (different serogroup B infections)
occurred in the study population but these were in the spring term.
Statistical analysis
Questionnaire data were scanned with Formic, an electronic scanning
package8 and stored in Microsoft Access (version 2.0).
We used Epi.Info (version 6.04) for X2 and Fisher's
exact tests and spss for Windows (version 8) for multiple logistic
regression analysis. Data collected at the time of the first pharyngeal
swab were used to determine risk factors for initial carriage through
multiple logistic regression. Subsequently, further analyses of
risk factors for acquisition during the first term were performed
with data from the repeat questionnaires and included only those
students whose pharyngeal swab was negative in October.
Results
Overall, 2507 first year students attended the university's health
centre in the first week of the first term, of whom 2453 (97.8%)
agreed to participate. A rapid increase in carriage of N meningitidis
occurred during the first week (table 1). Date of swabbing, type
of hall, active and passive smoking, and intimate kissing were all
independent risk factors for meningococcal carriage during the first
week (table 2). In November, 714 of the 939 eligible students (76.0%)
were reinvestigated for meningococcal carriage and social behaviour.
We could not process 172 swabs owing to a problem with an incubator,
leaving 542 students (57.7%). In December, 653 of 933 students (70.0%)
in catered halls who had participated in the first round were reinvestigated
along with 149 of 358 (42%) students in the self catered hall. The
figure shows the schedule of swabbing from October to December.
Schedule
for pharyngeal swabbing, October to December 1997
| Table
1 - Carriage rate of Neisseria meningitidis during first
week of term, 1997 |
| |
No
of students |
|
| |
 |
|
| Date |
Swabbed |
Positive
for N meningitidis |
%
carriage rate (95% CI) |
|
| 30
September |
825 |
57 |
6.9
(5.3 to 8.9) |
|
| 1
October |
669 |
75 |
11.2
(8.9 to 13.6) |
|
| 2
October |
691 |
131 |
19.0
(16.0 to 21.9) |
|
| 3
October |
268 |
62 |
23.1
(18.1 to 28.2) |
x2
for linear trend = 74.
P<0.0001. |
|
| Table
2 - Risk factors for carriage of Neisseria meningitidis
during first week of term |
| Exposure |
Odds
ratio (95% CI) |
P
value |
| Day
of throat swab |
| Tuesday |
Reference |
|
| Wednesday |
1.74
(1.2 to 2.5) |
0.003 |
| Thursday |
2.99
(2.1 to 4.2) |
0.0001 |
| Friday |
4.05
(2.7 to 6.1) |
0.0001 |
| Type
of hall |
| Mixed
sex |
Reference |
|
| Male
only |
1.16
(0.8 to 1.6) |
0.3 |
| Female
only |
0.77
(0.5 to 1.3) |
0.4 |
| Off
campus |
0.64
(0.5 to 0.9) |
0.01 |
| Passive
smoking (days) |
| 0.2 |
Reference |
|
| 3.4 |
1.26
(0.8 to 1.9) |
0.3 |
| 5.6 |
1.60
(1.1 to 2.4) |
0.2 |
| 7 |
2.03
(1.3 to 3.1) |
0.001 |
| Smoker |
2.40
(1.6 to 3.7) |
0.0001 |
| No
of people kissed |
| 0 |
Reference |
|
| 1 |
1.41
(1.1 to 1.8) |
0.01 |
| >or=2 |
1.19
(0.7 to 2.0) |
0.5 |
| Each
risk adjusted for all other variables in table and antibiotic
use in previous month. |
The carriage rate of N meningitidis increased during the first
term (table 3). Six weeks after widespread treatment with ciprofloxacin
the carriage rate in the excluded hall was 40 of 142 (28.2%, 95%
confidence interval 20.8 to 35.6). The carriage rate of group C
meningococci among students resident in catered halls was 0.5% (0.2
to 1.0) in October, 1.9% (0.9 to 3.9) in November, and 3.1% (2.0
to 4.4) in December. At some point during the first term, 349 students
with initially negative pharyngeal swabs in October acquired meningococci.
Table 4 shows the independently significant factors associated with
meningococcal acquisition during this period. Overall, 300 of the
325 index strains isolated in October (92%) and 333 of the 349 strains
acquired by students during the autumn term (95%) were fully typed.
Table 5 shows the full typ. ing of the commonest isolates in the
index round and the acquired strains. Although non.C strains predomi.
nated on October 1997 (mainly B and non.groupable meningococci),
C:2a:P1.5 (P1.2) was the commonest strain acquired (21 of 333) during
the first term (X2 10.8, P = 0.001.)
| Table
3 - Carriage of Neisseria meningitidis by study month
and hall status among first year undergraduate university students,
1997 |
| |
|
No
of students |
|
| Month |
Type
of hall |
Swabbed |
Positive
for N meningitidis |
%
carriage rate (95% CI) |
| October |
Catered |
1872 |
261 |
13.9
(12.4 to 15.1) |
| October |
Self
catered |
358 |
37 |
10.3
(7.2 to 13.5) |
| November |
Catered* |
542 |
168 |
31.0
(27.1 to 34.9) |
| December |
Catered* |
653 |
223 |
34.2
(30.5 to 37.8) |
| December |
Catered
|
142 |
40 |
28.2
(20.8 to 35.6) |
| December |
Self
catered |
149 |
36 |
24.2
(17.3 to 31.0) |
*Excludes
results where 30% of students received ciprofloxacin.
30% of students given ciprofloxacin in late October. |
| Table
4 - Risk factors for acquisition of Neisseria meningitidis
during first term |
| Exposure |
Odds
ratio (95% CI) |
P
value |
| Sex |
|
|
| Female |
Reference |
|
| Male |
1.61
(1.2 to 2.2) |
0.003 |
| Passive
smoking (days) |
|
|
| 0-2 |
Reference |
|
| 3-7 |
1.21
(0.9 to 1.7) |
0.2 |
| Smoker |
1.6
(1.0 to 2.6) |
0.05 |
| Weekly
visits to hall bar |
|
|
| 0 |
Reference |
|
| 1-4 |
1.74
(1.1 to 2.8) |
0.03 |
| >=5 |
2.71
(1.5 to 4.8) |
0.0005 |
| Type
of hall |
|
|
| Mixed
sex |
Reference |
|
| Male
only |
0.71
(0.5 to 1.0) |
0.05 |
| Female
only |
0.52
(0.3 to 0.9) |
0.03 |
| Self
catered |
0.73
(0.5 to 1.2) |
0.2 |
| Visited
night club |
|
|
| No |
Reference |
|
| Yes |
1.25
(1.0 to 1.6) |
0.05 |
| No
of people kissed |
|
|
| 0 |
Reference |
|
| 1 |
1.37
(1.0 to 1.8) |
0.04 |
| >=2 |
1.37
(0.9 to 2.1) |
0.1 |
| Each
risk adjusted for all other variables in table and antibiotic
use in previous month. |
Discussion
Our results show that meningococcal carriage increases rapidly
among university students in the first month of the academic year
and that much of this increase probably occurs during the first
week. Rapid acquisition rates have previously been found among military
recruits; however, these studies were generally smaller and fundamental
differences in sleeping arrangements existed compared with students.
9.11 Several explanations for the rapid increase we observed can
probably be discounted. The first was an improvement in swabbing
techniques over the first week of the study. Although we were unable
to identify the person who took each swab, most were taken by one
person (KRN) with considerable experience.12 KRN also
supervised the technique of the other operators. Furthermore, on
each day during the first week, different operators assisted with
swabbing in the morning and afternoon sessions yet there were no
differences between morning and afternoon carriage rates on any
day. We therefore believe that reliability was high between operators
taking the swabs. The alternative explanation is that students who
were more likely to be carrying meningococci on arrival at university
were recruited later in the week. This seems unlikely as over 99%
of students attended at their allotted time, and it seems unlikely
that any systematic bias would have been introduced by choice of
degree course. Furthermore, students are not allocated to halls
of residence by course or faculty groups. The association of carriage
with markers of social mixing also supports a causal link with acquisition
after arrival at university.
Our main finding was a rapid increase in meningococcal carriage
from 8% to 23% during the first week. Although the initial carriage
rate was surprisingly low (8%), this finding has now been replicated
by a subsequent study performed in October 1999 with a different
population of students, who had pharyngeal swabs taken both on arrival
and one week later (data available on request). Therefore, we do
not believe the initially low carriage rate to be artefactual and
speculate that clearance of meningococci occurs during the summer
holiday between leaving school and starting university. This may
arise from dispersal of the sixth form group, resulting in lower
rates of recolonisation.
We also noted that during the first week carriage was higher in
catered halls. This agrees with a previous study, which identified
an increased risk of invasive meningococcal disease at universities
offering com. paratively more accommodation in catered halls. We
speculate that this may be due to fundamental differences in the
physical structure and pattern of social interaction between catered
and self catered halls at Nottingham University.
Risk factors for carriage
In our regression analysis, we identified active and passive smoking
and intimate kissing as risk factors for carriage. These have been
previously shown by other investigators.13 In addition,
we noted that students living off campus were less likely to be
carriers, which is also consistent with the theory of social mixing.
| Table
5 - Typing data from 300 carriers in October 1997 and 333
strains acquired in first term in students previously negative
for N meningitidis |
| Full
typing details |
No
(%) of carriers |
No
(%) of newly acquired strains |
| C:2a:P1.5,2
(P1.5) |
3
(1.0) |
21
(6.3) |
| W135:NT:P1.3,6 |
8
(2.7) |
19
(5.7) |
| NG:NT:P1.16 |
7
(2.3) |
17
(5.1) |
| NG:NT:NT |
16
(5.3) |
11
(3.3) |
| NG:NT:P1.3,6 |
7
(2.3) |
10
(3.0) |
| NG:NT:P1.5 |
16
(5.3) |
5
(1.5) |
| B:NT:P1.15 |
10
(3.3) |
6
(1.8) |
| B:NT:NT |
9
(3.0) |
1
(0.3) |
| B:NT:P1.9 |
9
(3.0) |
0
(0) |
| NG:NT:P1.15 |
9
(3.0) |
5
(1.5) |
| 29E:NT:P1.5,2 |
8
(2.7) |
6
(1.8) |
| NG:15:P1.6 |
8
(2.7) |
5
(1.5) |
| NG:4:NT |
7
(2.3) |
3
(0.9) |
| X:21:P1.16 |
0
(0) |
9
(2.7) |
| Y:NT:P1.5,2 |
6
(2.0) |
9
(2.7) |
| B:1:P1.13 |
0
(0) |
7
(2.1) |
| NG=not
grouped; NT=not typed. Group C versus other strains Ç
2=10.8 (1df). P=0.001. |
Risk factors for acquisition
Although all of the students in this analysis had an initial negative
pharyngeal swab result, it is possible that some students were incorrectly
identified as non. carriers during the first week of the study.
This type of misclassification bias, inevitable in this type of
study, will have had the effect of weakening any associations detected.
For the same reason, our estimates of the prevalence of carriage
should also be regarded as conservative. Nevertheless, we identified
male sex, active smoking, visits to hall bars and nightclubs, intimate
kissing, and mixed sex halls as risk factors for acquisition. Most
of these factors have been previously identified in carriage and
outbreak studies, 14.16 but few have been addressed the risk of
acquiring carriage.11 The lower rate of acquisition seen
in female only halls probably reflects different patterns of social
behaviour.
As expected, NG (not grouped) and serogroup B strains predominated
in all swabs. Whereas carriage rates of serogroup C meningococci
are significantly lower than for serogroup B or non.groupable meningococci,
and typically less than 1% even in outbreaks,12,14,17
our study found serogroup carriage rates of 3% by December. This
level increases the risk of outbreaks. Serogroup C disease is of
particular importance as it is preventable by vaccine and has previously
been linked to large clusters of disease among university students.5,6
The large comparative increase in the C:2a:P1.5 (P1.2) strains
is noteworthy. This strain is known to be virulent and has been
implicated in several major out. Breaks3,5,6; it represented
6.3% of all acquired strains in our study but only 1% of index strains
identified in October. This illustrates the ability of highly virulent
clones to transmit readily among students. Indeed the preferential
transmission of this strain from a low base. line carriage rate
may explain the 3.5 week delay usually observed between the start
of university term and the peak incidence of cases and outbreaks.
During the beginning of university terms there is a rapid spread
of meningococci in first year students, which is probably associated
with social mixing, especially in catered halls. Our findings support
the recent introduction of meningococcal vaccination for university
students.
We thank Dr Jim Pearson for his advice on methods and statistics,
Dr Angela White and her colleagues in the Cripps Health Centre for
use of facilities and general support, Keith Ashford for culturing
the meningococci, and the Public Health Laboratory Service's meningococcal
reference unit in Manchester for serogrouping, serotyping, and serosubtyping
data.
Contributors: KRN initiated the study; he will act as guarantor
for the paper. KRN, JSN.V.T, and DAAA'A supervised the study. KRN,
JSN.V.T, NJ, KJ, RCBS, and RJM designed the study protocol. KRN
and NJ analysed the data. All investigators contributed to the final
paper.
Critical
appraisal: points to note
What does this paper tell you?
Firstly, that carriage of Neisseria meningitidis among
students at Nottingham seems to increase rapidly during the
first week of term. Students swabbed on day 4 were more likely
to be carriers than students swabbed on day 1. Secondly, the
study looks into the factors associated with being a carrier.
Several look promising: being male, a smoker, going to the
hall bar or to night clubs regularly, and living on campus.
Kissing may also be a risk factor, but it's odd that kissing
one person is linked to carrying meningococcus when kissing
more than two people is not (tables 2 and 4). Remember that
data from cross sectional surveys (table 2) can only suggest
that two things are linked, not that one causes the other.
It's always tempting to apply causality to associations in
cross sectional surveys, especially when they are plausible
(kissing spreads diseases, for example). Don't do it.
Should you believe it?
These authors make deductions about the spread of meningoccocus,
particularly in the first week of term, from a series of snapshots.
Each snapshot is of a different group of students, and the
later ones are incomplete to a greater or lesser extent. There
is plenty of potential for bias in this design, some of which
is discussed by the authors. The best way to answer the question,
"How far and how fast does meningococcus spread through students?"
is to take a complete cohort-all entrants one October, say-and
swab them at regular intervals. Anything less, and the evidence
is circumstantial.
Unfortunately, as is often the case, more robust designs
are also more time consuming and impractical. These authors
took the pragmatic approach and worked with what they had-groups
of students scheduled for successive days-but strengthened
their study by reswabbing students in halls later in the term
to give them some prospective, or longitudinal, data. This
strategy gave them 349 students who were initially negative
but acquired carriage during their first term. The longitudinal
part of the study is incomplete because of an accident at
the lab, so any conclusions drawn from it should be cautious,
such as risk factors for acquiring carriage. Incomplete data
should always ring alarm bells; you can never tell what the
missing data might have done to the results (although you
can often make an educated guess).
Alison Tonks, studentBMJ
Funding: Meningitis Research Foundation, Bristol.
Competing interests: None declared.
Keith R Neal, senior lecturer,
Email: keith.neal@nott.ac.uk
Jonathan S Nguyen-Van-Tam, senior lecturer,
Nicholas Jeffrey, medical student,
Richard J Madeley, professor,
Katy Job, medical student, Department of Public Health Medicine and Epidemiology, University of Nottingham, Queen's Medical Centre, Nottingham NG7 2UH
Richard C Slack, senior lecturer,
Kamel Ait-Tahar, PhD student,
Dlawer A A Ala'Aldeen, reader, Meningococcal Research Group, Division of Microbiology, Queen's Medical Centre
Martin C J Wale, regional epidemiologist, Communicable Disease Surveillance Centre Trent, Queen's Medical Centre
studentBMJ 2000;08:131-174 May ISSN 0966-6494
- Hubert B, Caugant DA. Recent changes in meningococcal disease in Europe. Euro Surveillance 1997;2: 69.71.
- Whalen CM, Hockin JC, Ryan A, Ashton F. The changing epidemiology of invasive meningococcal disease in Canada, 1985 through 1992. Emergence of a virulent clone of Neisseria meningitidis. JAMA 1995;273:390.4.
- Jackson LA, Schuchat A, Reeves MW, Wenger JD. Serogroup C meningococcal outbreaks in the United States. An emerging threat. JAMA 1995;273:383.9.
- Neal KR, Nguyen.Van.Tam JS, Monk P, O'Brien SJ, Stuart J, Ramsay M. Invasive meningococcal disease among university undergraduates: association with catered halls of residence. Epidemiol Infect 1999;122:351.8.
- Anon. Outbreak of meningococcal disease in students in Cardiff. Commun Disease Rep 1996;6:425.
- Anon. Clusters of meningococcal disease in university students. Commun Disease Rep 1997;7:393.
- Jackson LA, Schuchat A, Gorsky RD, Wenger JD. Should college students be vaccinated against meningococcal disease? A cost.benefit analysis. Am J Public Health 1995;85:843.5.
- Formic, an integrated questionnaire design and scanning package. http://www.formic.co.uk. (Accessed 7 September 1999.)
- Anderson J, Berthelson L, Jensen BB, Lind I. Dynamics of the meningococcal carrier state and characteristics of the carrier strains: a longitudinal study within three cohorts of military recruits. Epidemiol Infect 1998;121:85.94.
- Devine LF, Pierce WE, Floyd TM, Rhode SL, Edwards EA, Siess EE, et al. Evaluation of group C meningococcal polysaccharide vaccine in marine recruits, San Diego, California. Am J Epidemiol 1970;92:25.32.
- Melton LJ, Edwards EA, Devine EA, Devine LF. Differences between the sexes in the nasopharyngeal carriage of Neisseria meningitidis. Am J Epidemiol 1977;106:215.21.
- Neal KR, Irwin DJ, Davies S, Kaczmarski EB, Wale MCJ. Sustained reduction in the carriage of Neisseria meningitidis as a result of a community meningococcal disease control programme. Epidemiol Infect 1998;121:487.93.
- Stuart JM, Cartwright KAV, Robinson PM, Noah ND. Effect of smoking on meningococcal carriage. Lancet 1989;ii:723.5.
- Rønne T, Berthelesen L, Buhl LH, Lind I. Comparative studies on pharnygeal carriage of Neisseria meningitidis during a localized outbreak of serogroup C meningococcal disease. Scand J Infect Dis 1993;25:331.9.
- Imrey PB, Jackson LA, Ludwinski PH, England III AC, Fella GA, Fox BC, et al. Outbreak of serogroup C meningococcal disease associated with campus bar patronage. Am J Epidemiol 1996;143:624.30.
- Jelfs J, Jaluludin B, Munro R, Patel R, Kerr M, Daley D, et al. A cluster of meningococcal disease in western Sydney, Australia initially associated with a nightclub. Epidemiol Infect 1998;120:263.70.
- Le Saux N, Ashton F, Rahman F, Ryan A, Ellis E, Tamblyn S, et al. Carriage of Neisseria species in communities with different rates of meningococcal disease. Can J Infect Dis 1992;3:60.4.
