Emergence of W135 meningococcal meningitis in Ghana


Gerd Pluschke (corresponding author), Julia Leimkugel, Jean-Pierre Dangy, Sebastien Gagneux and Tom Smith, Swiss Tropical Institute, Socinstr. 57, 4002 Basel, Switzerland. E-mail: Gerd.Pluschke@unibas.ch
Abudulai Adams Forgor, Abraham Hodgson and Akalifa Bugri, Navrongo Health Research Centre, Ministry of Health, Navrongo, Ghana.


Neisseria meningitidis serogroup W135, well known for a long time as a cause of isolated cases of meningococcal meningitis, has recently increasingly been associated with disease outbreaks of considerable magnitude. Burkina Faso was hit by W135 epidemics in the dry seasons of 2002–2004, but only four W135 meningitis cases were recorded between February 2003 and March 2004 in adjoining Ghana. This reconfirms previous findings that bottlenecks exist in the spreading of new epidemic N. meningitidis clones within the meningitis belt of sub-Saharan Africa. Of the four Ghanaian W135 meningitis patients one died and three survived, of whom one had profound neurosensory hearing loss and speech impairment. All four disease isolates were sensitive to penicillin G, chloramphenicol, ciprofloxacin and cefotaxime and had the multi-locus sequence type (ST) 11, which is the major ST of the ET-37 clonal complex. Pulsed-field gel electrophoresis (PFGE) profiles of the Ghanaian disease isolates and recent epidemic isolates from Burkina Faso were largely identical. We conducted meningococcal colonization surveys in the home communities of three of the patients and in the Kassena Nankana District located at the border to Burkina Faso. W135 carriage rates ranged between 0% and 17.5%. When three consecutive surveys were conducted in the patient community with the highest carrier rate, persistence of W135 colonization over a period of 1 year was observed. Differences in PFGE profiles of carrier isolates taken at different times in the same patient community were indicative of rapid microevolution of the W135 bacteria, emphasizing the need for innovative fine typing methods to reveal the relationship between W135 isolates.


Epidemic meningococcal disease has occurred in the meningitis belt of sub-Saharan Africa for approximately 100 years (Greenwood 1999). Historically the epidemics have been primarily caused by Neisseria meningitidis serogroup A. Serogroup W135 meningococci, identified in 1968 (Evans et al. 1968) and first described in Africa in 1982 (Denis et al. 1982), were initially considered to be a rare cause of invasive disease. However, two W135 meningitis outbreaks coinciding with pilgrimage seasons for Hajj in 2000 and 2001 (Taha et al. 2000; Lingappa et al. 2003) were followed by the first large scale epidemic in Burkina Faso in 2002 (Decosas & Koama 2002; Taha et al. 2002). Since then, each year Burkina Faso has been hit by mixed meningitis epidemics caused by W135 and A meningococci. In Saudi Arabia W135 meningococci were responsible for 13% of all meningococcal disease between 1995 and 1999 and have been present to a notable degree at least since 1990 (Lingappa et al. 2003). From 2002 onwards vaccination with the quadrivalent meningococcal polysaccharide vaccine (A/C/Y/W135) therefore became a visa requirement to participate in the Hajj (Wilder-Smith 2003). Even before the outbreaks in 2000 the danger of W135 meningitis epidemics in Africa was recognized (Kwara et al. 1998).

The Hajj outbreaks probably led to the expansion of a particular W135 clone within the electrophoretic type-37 (ET-37) complex (Popovic et al. 2000; Mayer et al. 2002). A high acquisition rate of W135 meningococci (15–17%) in pilgrims has been reported (Wilder-Smith et al. 2003). Throughout the world these carriers have transmitted Hajj-related W135 bacteria after returning home (Aguilera et al. 2002; Hahne et al. 2002; Wilder-Smith et al. 2003). Related W135 strains also belonging to the ET-37 complex have been circulating worldwide since at least 1970 (Mayer et al. 2002) and currently both the Hajj-related epidemic strain and Hajj-unrelated local W135 strains seem to be responsible for sporadic W135 cases worldwide (Hahne et al. 2002; Taha et al. 2004). Genetic drift of the Hajj-related strain (Hahne et al. 2002) complicates the analysis of the relationship between W135 isolates by standard typing techniques, such as pulsed-field gel electrophoresis (PFGE) and multi-locus sequence typing (MLST) considerably.

In spite of its border with Burkina Faso, no outbreak of W135 meningococcal meningitis has so far occurred in Ghana. Here, we describe properties of four W135 strains isolated between February 2003 and March 2004 from the cerebrospinal fluid (CSF) of Ghanaian meningitis patients and provide evidence for spreading and rapid microevolution of the causative W135 meningococci.

Materials and methods

Disease isolates

CSF samples were taken in hospitals for diagnostic purposes, latex agglutination was performed and the causative agents were isolated by culture using standard microbiological techniques. Bacterial isolates of all four Ghanaian W135 cases were transferred for further analysis to the Navrongo Health Research Centre, where serological grouping was reconfirmed by PCR. Reference isolates of W135 meningococci were obtained from M. Achtman, Berlin (isolated in Mecca, 2000, strains Z9230 and Z9232), and D. Caugant, Oslo [isolated during the outbreaks in Burkina Faso of 2001 (BF01/01, BF24/01), 2002 (BF06/02, BF67/02) and 2003 (BF01/03)].

Carrier isolates

In three of the affected communities and two control communities throat swabs were analysed for colonization with N. meningitidis and Neisseria lactamica. Community K1 is a small village located in a rural setting directly on the main truck road between the south and the north of Ghana (Figure 1). Nearly the whole population of the village participated in the study. In December 2003 a control community located 2 km away from K1 was included. Communities B1 and B2 are located in Bolgatanga, the Upper East Regional capital. Here throat swabs were taken from the affected and the closest neighbouring compounds (including the majority of the about 30 inhabitants per compound).

Figure 1.

Map of Ghana showing the location of home communities of W135 meningitis patients and the date of disease onset. The sample time points were as follows: Community K1: (I) April 2003, (II) December 2003 (including control community), (III) April 2004. Community B1: December 2003, Community B2: March 2004. The home community of patient A1 could not be identified and has not been sampled.

After obtaining informed consent, throat swabs were taken and directly plated onto Thayer Martin Agar. The plates were incubated at 37 °C within 8 h after sampling for 24–48 h. Two colonies with neisserial morphology were sub-cultured from each plate. N. meningitidis and N. lactamica colonies were identified as previously described (Gagneux et al. 2002) by standard bacteriological methods. Ethical clearance was obtained from the responsible institutional and national ethical approval committees.

Characterization of bacterial isolates

Meningococcal isolates were serogrouped with serogroup-specific antisera (Difco). Results were reconfirmed by PCR (Orvelid et al. 1999; Taha et al. 2000; Bennett et al. 2004). All W135 isolates were analysed by PFGE after digestion with NheI as previously described (Morelli et al. 1997). All disease isolates were tested for resistance to penicillin G, chloramphenicol, cefotaxime, and ciprofloxacin with E-test strips (Isenberg 1998) using the NCCLS breakpoints. Selected strains were analysed by MLST. DNA extraction (Vela Coral et al. 2001), PCR (Maiden et al. 1998) and sequencing of PCR products with an ABI Prism 310 Genetic Analysis System were performed according to standard protocols on the MLST homepage (http://pubmlst.org/neisseria/). Allelic profiles were analysed using applications available on the MLST homepage.


Characterization of W135 disease isolates from Ghana

Between February 2003 and March 2004, four cases of W135 meningococcal meningitis were reported by the regional hospitals in Tamale, Bolgatanga and the Korle Bu Teaching hospital (Accra). Patients came from the centre, the south or the north of the country (Figure 1) and were 3–17 years old (Table 1). One patient died, and of the three survivors one had profound sequelae.

Table 1.  W135 cases reported to the Ghanaian disease control authorities in 2003 and 2004
Paitient IDTime of disease onsetVillage/cityRegionAge (years)SexOutcomeSequelae
  1. * Profound neurosensory hearing loss, speech impairment, transient ataxia, hyperactive left knee and Achilles reflexes.

  2. † Arthritis of the knee joints and occasional episodes of brief startling attacks during the first week after discharge but stopped thereafter.

K1February 2003KpalkpalgbeniBrong Ahafo3MaleSurvivedMultiple*
A1April 2003AccraGreater Accra4MaleDiedUnknown
B1August 2003BolgatangaUpper east17FemaleSurvivedNone
B2March 2004BolgatangaUpper east3MaleSurvivedMultiple†

All four disease isolates were sensitive to penicillin G, chloramphenicol, ciprofloxacin and cefotaxime. PFGE profiles of all four Ghanaian disease isolates were compared with disease isolates from the Hajj outbreak in Mecca 2000 and from Burkina Faso between 2001 and 2003. The Burkina strains isolated in 2001 and 2002 showed identical profiles (shown for strain BF67/02, Figure 2, lane 5, profile C), whereas the 2003 isolate appeared to be very closely related (Figure 2, lane 6, BF01/03, profile D). Profiles of the Ghanaian disease isolates were largely identical (Figure 2, lanes 7–10, profile D) and indistinguishable from that of the 2003 strain from Burkina Faso (Figure 2, lane 6). The reference disease isolates from the Hajj outbreak in 2000 had a distinct but related PFGE profile (shown for strain N11421, Figure 2, lane 4, profile B; both strains had an identical profile). All four Ghanaian disease isolates had the multi-locus sequence type (ST) 11, which is the major ST of the ET-37 clonal complex.

Figure 2.

PFGE profile of W135 carrier and disease isolates (lane: strain No.; origin) Indicated on the gel are the different band profiles of the W135 strains (A–G). 1: MW marker, 2, 3: N1621, N1622, KND 1998, carriage; 4: N1421 (Z9230), Mecca 2000, reference strains; 5: N1627 (BF67/02), Burkina Faso 2002, case; 6: N1628 (BF01/03), Burkina Faso 2003, case; 7: N1681 Ghana 2003, patient K1; 8: N1682, Ghana, 2003, patient A1 9: N1683, Ghana 2003, patient B1; 10: N1846, Ghana 2004, patient B2; 11, 12: N1485, N1487, community K1, April 03, carriage; 13, 14, 15: N1633, N1640, N1636, community K1, Dec 03, carriage; 16, 17: N1848, N1857, community B2, March 04, carriage; 18, 19, 20: N1951, N1953, N1959, community K1, April 03, carriage; 21, 22: N1888, N1903, KND, April 04, carriage; 23: MW marker.

W135 colonization in patient communities and clonal diversity of bacteria

Three consecutive N. meningitidis colonization surveys were performed in community K1, the first 6 weeks after the emergence of the case, in February 2003. In April 2003, 17.5% of 103 inhabitants (about 90% of the total population of the village) were colonized with W135 meningococci. Thereafter, the W135 colonization rate declined to 13% in December 2003 and to 3% in April 2004. In addition, a few carriers of other meningococci were found. N. lactamica colonization rates were between 3% and 8%. N. meningitidis A, X and Y, but no W135 carriers were found in a neighbouring control community included in the December 2003 survey (Table 2).

Table 2.  Carriage of different serogroups of Neisseria meningitidis and of Neisseria lactamica in home communities of three W135 meningococcal meningitis patient and in a neighbouring control community
Time of surveyCommunity (patient ID)Volunteers swabbed (n)Colonization rate % (n)
Neisseria lactamicaNeisseria meningitidisW135AXYNG*
  1. * Non-groupable (NG) strains were negative both in serological tests and in serogroup A and W135 specific PCR analyses.

April 2003Home (K1)1037.8 (8)24.3 (25)17.5 (18)2 (2)01 (1)3 (3)
December 2003Home (K1)1003 (3)15 (15)13 (13)001 (1)1 (1)
April 2004Home (K1)965.2 (5)8.3 (8)3.1 (3)1 (1)01 (1)3 (3)
December 2003Control (K1)1001 (1)9 (9)02 (2)2 (2)2 (2)3 (3)
December 2003Home (B1)1108.2 (9)3.6 (4)0001 (1)3 (3)
April 2004Home (B2)1004 (4)7 (7)2 (2)02 (2)02 (2)

Cumulated data from all three surveys conducted in community K1 were used to analyse the age distribution of colonization with W135 meningococci in comparison to other serogroups found and to N. lactamica (Table 3). Logistic regression, including random effects to allow for repeated assessment of the same individuals, indicated that the ratio of carriage prevalence of N. meningitidis to that of N. lactamica increased with age (χ2 = 7.6, 1 degree of freedom, P = 0.006), but there was no significant age trend in the ratio of W135 to other N. meningitidis (χ2 0.8, 1 d.f. P = 0.4).

Table 3.  Age distribution of colonization with Neisseria lactamica and W135 and non-W135 Neisseria meningitidis in the patient home community K1 (cumulated data from all three surveys)
Age group (years)Frequency of colonization
Neisseria meningitidis serogroup W1351/6 (16.7%)5/78 (6.4%)7/44 (15.9%)7/37 (18.9%)6/24 (25.0%)8/93 (8.1%)0/17 (0%)
Non-W135 N. meningitidis0/6 (0%)1/78 (1.3%)0/44 (0%)4/37 (10.8%)3/24 (12.5%)5/93 (5.3%)1/17 (5.9%)
Neisseria lactamica2/6 (33.3%)9/78 (11.5%)1/44 (2.3%)1/37 (2.7%)0/24 (0%)0/93 (0%)1/17 (5.9%)

All isolates from the 18 W135 carriers in community K1 in April 2003 revealed identical PFGE profiles (shown for strains N1485 and 1487, Figure 2, lanes 11 and 12, profile D), indistinguishable from those of the Ghanaian disease isolates (Figure 2 lanes 7–10). However, some genetic diversification became apparent in the December 2003 colonization survey. While the isolates of nine (of thirteen) W135- carriers revealed the original band profile (data not shown), the isolates from the other four exhibited three new variant profiles (Figure 2, lanes 13–15, profiles E, F, G). Two of the three variant PFGE profiles, but not the original profile, were found again in the last colonization survey in April 04 (Figure 2, lanes 18–20, profiles E and G).

In community B1 and B2 only one colonization survey was performed, 3 months and 3 weeks, respectively, after the emergence of the case. While no W135 meningococci were found in community B1, in community B2 W135 isolates of two carriers were obtained with the same PFGE profile as the Ghanaian disease isolates (Figure 2, lane 16, profile D, Table 2). In addition, from one of them a variant strain was isolated with a PFGE profile (Figure 2, lane 17, profile F) identical to a variant profile found in colonization isolates from community K1.

W135 colonization in a long-term colonization survey in northern Ghana

Within the framework of a longitudinal N. meningitidis colonization and disease study in the Kassena Nankana District (KND) of northern Ghana (Gagneux et al. 2000, 2002), no W135 meningococcal meningitis case was recorded between 1998 and 2004. During these 7 years of twice yearly colonization surveys only single carriers of W135 meningococci were identified in 1998 (1/300 in April and 1/299 in November 1998) (Gagneux et al. 2000). However, in April 2004 a W135 colonization rate of 0.9% (3/350) was found, with isolates showing the same PFGE profile (Figure 2, lanes 21 and 22, profile D) as the Ghanaian disease isolates (Figure 2, lanes 7–10). Profiles of the two 1998 carrier isolates (Figure 2, lanes 2 and 3, profile A), were identical to each other but distinct from all other profiles observed in this study.


In spite of the consecutive W135 epidemics in Burkina Faso in 2002–2004, no major outbreak of W135 disease has been observed so far in Ghana, demonstrating that bottlenecks exist for the spreading of epidemic strains within the meningitis belt, as already described for serogroup A meningococci (Achtman 1995). The four isolated Ghanaian cases described in this paper have probably only been reported because of intensive national surveillance and awareness. W135 strains belonging to the ET-37 complex were present in Ghana before the Mecca outbreak (Gagneux et al. 2000) and sporadic W135 cases may easily have remained undetected before the year 2000.

PFGE analysis demonstrates that the four Ghanaian W135 meningitis isolates were closely related to recent disease isolates from Burkina Faso, indicating that these meningitis cases were caused by epidemic-related strains and not by local strains of the ET-37 complex. At least in the north of Ghana colonization with the Burkina Faso epidemic-related strain is detectable. While visitors from Burkina Faso are frequently met in the border communities B1 and B2, it is not possible to guess the origin of the disease causing W135 strain of patient A1 living in Accra. In the case of community K1, located in the middle of Ghana, contact to nomads may have been the source of the W135 bacteria, as a part of a neighbouring community frequently moves to Burkina Faso and back.

W135 carriage rates of healthy contacts in the three home communities of W135 meningitis patients, were very different. W135 carriers were found in the home communities (K1 and B2) of the index cases aged 3 years but not in B1, the home community of the 17-year-old patient. Age of the patients may play a role, as suggested by findings of a study carried out during a serogroup C outbreak in Brazil, where contact carriage rates were highest in households, where the index case was an infant (Cartwright 1995). Carriage rates of outbreak strains tend to be higher in closed or partially closed communities than in open communities (Cartwright 1995). The rural community K1 has the features of a semi-closed community, where inhabitants lived very closely together and shared all living activities, while the urban communities B1 and B2 were much more open and loose. This may explain why the highest (18%) carriage rate was observed in community K1. The age distribution of W135 colonization was not unusual, as the pattern observed in community K1 was characteristic for meningococci in general (Cartwright 1995).

Changes of the PFGE profile of colonization isolates with time demonstrate that microevolution of W135 may be rapid. N. meningitidis is a naturally transformable species and there is evidence that microevolution is driven more frequently by recombination than by mutation. The observed genetic drift can make it very difficult to distinguish between epidemic-related and local W135 strains belonging to the same ET-37 complex and to prove epidemic spread of a particular clone. While available techniques are suitable to analyse the global population structure of other meningococcal serogroups (Lingappa et al. 2003), new approaches are required for studying the molecular epidemiology of N. meningitidis W135.

An affordable vaccine against W135 meningococci (e.g., a trivalent groups A, C, and W135 polysaccharide vaccine) is now available and has been successfully used to contain outbreaks of W135 meningitis in Burkina Faso (Ahmad 2004). As Burkina Faso epidemic-related W135 meningococci now seem to spread into Ghana, intense surveillance efforts at national and regional levels for timely detection of a potential W135 epidemic is an important issue in future years.


This study was supported by Grant GAT.0779-01476-GRT of the Meningitis Vaccine Project. We gratefully acknowledge the National Disease Control Unit of the Ghana Health Service and the National Public Health Reference Laboratory for releasing the CSF samples of the index case and of the case from Accra. Furthermore we are grateful to the laboratory personnel of the Bolgatanga Regional hospital for making available the other two CSF samples. We thank the CSM team of the NHRC for assistance during the study and the communities, health authorities (Upper East Regional Health Directorate, Bolgatanga Municipal Health Directorate, Kintampo District Health Directorate, Kintampo Health Research Centre, Navrongo Health Research Centre) as well as all study participants for the cooperation accorded us. Finally we acknowledge D. Caugant and M. Achtman for helpful comments on the manuscript and the supply with W135 Reference isolates.