Epidemiological aspects of the 1969 yellow fever epidemic in Nigeria

Bull. Org. mond. Sant} 1972, 46, 645-651 Bull. Wid Hlth Org. Epidemiological aspects of the 1969 yellow fever epidemic in Nigeria D. E. CAREY," G. E. KEMP,' J. M. TROUP,2 H. A. WHITE,2 E. A. SMITH,3 R. F. ADDY,4 A. L. M. D. FOM,4 J. PIFER,5 E. M. JONES,6 P. BRLS,7 & R. E. SHOPE 8 The Virus Research Laboratory of the University of Ibadan, Nigeria, was notified on 23 October 1969 that cases of suspected yellow fever had occurred in the Jos area. The diagnosis was confirmed by virus isolation and the existence of a widespread outbreak on the Jos Plateau and adjacent areas was established. This was the first recognized epidemic ofyellow fever in Nigeria since 1953. Between September and the end ofDecem- ber 1969, an estimated total of252 patients with yellow fever were hospitalized. The case- fatality ratio for hospitalized patients was approximately 40 %. The diagnosis of yellow fever was confirmed by virus isolation, serology, or pathology in 55 patients. It is estimated that up to 100 000 cases of yellow fever may have occurred during the epidemic. In late October 1969, the Virus Research Labora- tory of the University of Ibadan, Nigeria, was noti- fied that cases of suspected yellow fever had occurred on the Jos Plateau. The diagnosis was confirmed within a few days by both virus isolations and sero- logical tests. Further investigations established that there was an extensive outbreak involving the pla- teau and adjacent areas. The region has been described by Lee (1972). Epidemiological observa- tions are reported in this article. MATERIALS AND METHODS Virus isolation was carried out in infant mice by standard methods (Causey & Kemp, 1968). Comple- ment-fixation tests were performed on plastic plates (Weinbren, 1958). Serological diagnoses were based on the demonstration by complement fixation of a rise in the yellow fever antibody titre between 1 Virus Research Laboratory, University of Ibadan, Nigeria. This laboratory receives support from the Rocke- feller Foundation. ' Formerly at Evangel Hospital, Jos, Nigeria. (Deceased 1970.) ' Federal Ministry of Health, Lagos, Nigeria. 'Ministry of Health, Benue-Plateau State, Nigeria. 'Center for Disease Control, Atlanta, Ga., USA. Vom Christian Hospital, Vom, Nigeria. Medical Officer, Virus Diseases, World Health Organiza- tion, Geneva, Switzerland. ' Yale Arbovirus Research Unit, Yale University School of Medicine, New Haven, Conn., USA. paired sera, or the presence of yellow fever comple- ment-fixing antibody in one or more serum speci- mens. Yellow fever, West Nile, and Zika virus antigens were employed. Neutralization tests on survey sera were made in infant mice inoculated intracerebrally with a single virus dose mixed with an equal volume of serum. Virus doses ranged from 200 LD50 to 2 000 LDI50 and averaged 800 LD50. Results were recorded as positive if either 5 or 6 of the 6 inoculated mice survived. RESULTS Diagnosis Laboratory confirmation of the diagnosis was available for 55 patients: 15 by isolation of the virus from blood serum, 2 by virus isolation and comple- ment-fixation test, 12 by pathology, 1 by pathology and complement-fixation test, and 25 by complement- fixation test only.9 The first isolate of yellow fever virus at Ibadan (IbH 43913) was recovered from the serum of a 20-year-old man who had been ill for 4 days prior to the sampling of blood in Jos on 18 October 1969. After being kept for 6 days in the freezing compart- ' Three virus isolations and 3 pathological diagnoses were made by Dr W. G. C. Bearcroft, West African Council for Medical Research, Yaba, Nigeria. The remaining patho- logical diagnoses were reported by Professor G. M. Edington University of Ibadan, Nigeria. 2842 -645- D. E. CAREY ET AL. ment of a refrigerator, the specimen was sent by air on wet ice to Ibadan where it was received on 24 October and inoculated into 2 groups of infant mice. All the mice became ill from the eighth post- inoculation day onwards. The brain suspension of a sick mouse sacrificed on the eighth day was employed as complement-fixing antigen and tested against the following mouse immune ascitic fluids (MIAF): dengue 1 and 2, Banzi, Spondweni, Usutu, Zika, Uganda S, West Nile, multiple group B, and yellow fever. Reactions were noted only with the group B and yellow fever MIAF. This identification was confirmed in a baby mouse neutralization test in which IbH 43913 was neutralized by a yellow fever MIAF. Subsequent yellow fever isolates were identified on the basis of complement-fixation tests alone. Most virus isolations were made from persons seen in outpatient clinics who complained of fever, headache, epigastric pain, and myalgia. Virus was isolated from patients up to and including the fifth day of illness. Of the 25 patients in whom yellow fever was diag- nosed by the complement-fixation test alone, 13 showed a rise or fall in the yellow fever complement- fixing antibody titre between paired serum specimens, and 12 were accepted on the basis of the presence of yellow fever antibody in 1 or more specimens. Representative serological findings are presented in Table 1. Case 132 is believed to be an instance of yellow fever in a person with prior group B antibody. Hospitalized cases In order to obtain an estimate of the number of people hospitalized, records at the 7 hospitals that had the largest number of admissions (3 mission and 4 government hospitals) were examined for admissions classified as hepatitis, jaundice, and yellow fever from 1 June up to and including 31 December 1969. Patients with confirmed cancer or cirrhosis were omitted. Records of similar diag- noses in previous years were also recorded when they were available. During the last 7 months of 1969, 397 patients were hospitalized with the listed diagnoses and there were 123 deaths (Fig. 1). The first laboratory con- firmation of yellow fever, however, was in a patient whose illness began on about 7 September. At that time also the number of deaths started to increase. If this date is accepted as a conservative estimate of the beginning of the epidemic period, there were 307 hospital admissions and 116 deaths. Table 1. Complement-fixing antibody responses of yellow fever patients from Jos Plateau, 1969 Antigen a Serial Date of sampling Yellow West Zika no. Ylo et Zk fever Nile 106 18 October neg. neg. neg. (virus isolated) 18 December 32 b neg. neg. 119 27 October neg. 8 neg. (virus isolated) 7 November 512 128 8 132 18 October 512 512 64 24 October 512 512 32 290 25 November 512 neg. 64 291 25 November 128 neg. neg. 364 7 November 256 neg. neg. 8 December 32 neg. neg. aneg.== <1 :8. b Reciproval of serum titre. Infectious hepatitis is frequently seen in Nigeria, and it would be expected to continue at previous levels even during a yellow fever epidemic. A review of the records for 4 previous years at 6 plateau hospitals revealed that an average of 47 cases per year diagnosed as hepatitis or jaundice were admitted from 7 September to the end of the year. These 80 60 Deaths E 50 40 y30 -::> S c20- 10 J J A S 0 N Period ended Fig. 1. Admissions with diagnoses of yellow fever, hepatitis, or jaundice at 7 hospitals in Benue-Plateau and North-Eastern states, Nigeria, in June-December 1969. 646 EPIDEMIOLOGY OF A YELLOW FEVER EPIDEMIC IN NIGERIA 14 28 12 26 9 23 6 20 4 18 1 15 29 13 27 J J A S 0 N D Period ended Fig. 2. Admissions with diagnosis of hepatitis or jaundice at 6 hospitals, in Benue-Plateau and North- Eastern States, Nigeria, 1965-68 (average per year per 14-day period), and hepatitis, jaundice, and yellow fever admissions in 1969. admissions are plotted by 2-weekly periods in Fig. 2 and provide a basis for evaluating the situation in 1969; the increased number of admissions in 1969 is clearly seen. If infectious hepatitis did in fact occur in 1969 at a level similar to that seen in previ- ous years, then approximately 18% of the 307 damissions at plateau hospitals shown in Fig. I from 7 September 1969 could have been for hepa- titis. The rest of the admissions are presumed to represent cases of yellow fever, so that a final esti- mate of 252 hospitalized yellow fever patients is obtained. Deaths from hepatitis are uncommon and most of the fatal jaundice during this period can be attributed to yellow fever. Attributes of the epidemic As noted above, it was known from 7 September 1969 that cases of yellow fever were occurring. Fig. 2, however, reveals the presence of excess jaundice admissions from August of that year, and these cases may reflect early yellow fever activity. No cases of yellow fever are known to have occurred after December 1969, and although sporadic cases of jaundice continued, the virus studies that were made were negative for yellow fever. The peak of the epidemic came towards the end of an unusually heavy and prolonged rainy season, and at a period when minimum temperatures were rapidly falling to levels unfavourable for virus transmission. A striking feature of the epidemic was the appar- ent sparing of the residents of Jos town. Of 27 yellow fever patients admitted to Evangel Hospital in Jos, only 2 gave addresses in the town; the rest came from surrounding areas on the plateau. Fig. 3 Fig. 3. Jos Plateau and adjacent areas, showing the location of yellow fever cases. 647 648 D. E. CAREY ET AL. Table 2. Yellow fever serum survey at Jos Plateau, 1969 Complement fixation Neutralization test test Age Location (years) No. of sera Positive sera No.of sera Positive sera tested No. % tested No. % Gindiri 15-34 166 28 17 174 29 17 Du 0-14 77 3 14 75 5 7 15-34 63 7 1 1 64 9 1 4 35-54 26 2 8 26 5 1 9 55 9 0 0 10 0 0 Total for Du 175 12 7 175 19 11 shows the localities where cases are known to have occurred. Despite the presence of a potential vector (Aedes aegypti) in such urban centres as Ibadan and Lagos, and the dense traffic between these cities and the plateau, no cases of yellow fever were detected outside the epidemic area. The majority of those hospitalized were young adults between the ages of 15 and 34 years although the youngest patient recorded was a 4-year-old boy who had a fatal illness and confirmed cases were also seen in persons said to be over 50 years of age. Males outnumbered females by 4 to 1. As might be expected, because of the immunization require- ments for international travel, no member of the senior civil service or expatriate community was affected; however, most of these people lived in Jos town. The overall case-fatality ratio for yellow fever from 7 September for patients at the seven hospitals was 38 %. At Evangel and Vom Christian hospitals, where the data were perhaps most reliable, the ratios were 63% and 46 %, respectively. Serum surveys Surveys were carried out during the first week of November 1969 at Gindiri Teacher Training Col- lege,' 48 km south-east of Jos, and at Du, a village about 16 km south of Jos. Both villages had reported many cases of yellow fever and there were positive laboratory diagnoses from each area. At Gindiri only young adults were tested but at Du persons of all age groups were included. The results 1 The survey was made by courtesy of the Principal, Mr D. J. Williams. of complement-fixation and neutralization tests are presented in Table 2. At Gindiri, 17% of the sam- ples tested gave evidence, on the basis of positive complement-fixation tests, of recent yellow fever infections. The same percentage of samples, but not in all instances the same individuals, also had neutralizing antibody. At Du, 7% of those tested were found to have yellow fever complement-fixing antibody and 11% had neutralizing antibody. In most instances, yellow-fever-positive sera from both villages that were tested against all local group B antigens, reacted in complement-fixation tests exclusively with the yellow fever virus antigens at titres ranging from 1: 16 to 1: 64. Several sera were positive at a titre of 1: 8 only, i.e., at the lowest dilution employed. No significant sex differences were noted. Each student at Gindiri was asked about febrile illnesses within the previous 2 months. Among the 28 students whose sera were positive for com- plement-fixing antibody, 22 (79 %) reported a recent illness. In contrast, only 27% of the 138 students whose sera were negative for complement-fixing antibody reported an illness (Table 3). In conjunction with the testing of human sera, a, number of animal sera collected before the epi- demic in different areas of Nigeria were examined in yellow fever neutralization tests. All the 42 sera from hedgehogs (Atelerix albiventris) were nega- tive. Sera from 3 of 31 bushbabies (Galago sene- galensis), from 6 of 22 monkeys, and from 1 of 2 baboons (Papio anubis), were positive. The posi- tive monkey sera included samples from 4 Cercocebus torquatus and 2 Cercopithecus mona; 5 of these EPIDEMIOLOGY OF A YELLOW FEVER EPIDEMIC IN NIGERIA Table 3. Relation between yellow fever antibody and illness in students at Gindiri Teacher Training College. Complement- No. of liI Not ill fixing students antibody tested No. % No. % present 28 22 79 a 6 21 absent 138 37 27 101 73 total 166 59 35.5 107 64.5 a The proportion of students reporting illness was much higher among those with complement-fixing antibodies than among those without; the difference was statistically significant. 6 animals were collected in the Western State of Nigeria near the Dahomey border. The monkey and baboon sera were examined by haemagglutination-inhibition (HI) tests against a battery of group B antigens: yellow fever, West Nile, Wesselsbron, dengue 1 and 2, and Zika. Altogether, 20 of the 24 sera inhibited Zika antigen, in most instances at a serum dilution of at least 1: 80. All but 1 of the 7 sera positive in yellow fever neu- tralization tests also contained demonstrable yellow fever haemagglutination-inhibiting antibody, but titres were lower than those for Zika in 2 instances and equal in 3. Two of the 3 neutralization-test- positive Galago sera and 4 Galago sera that showed partial protection were available for HI testing. Only 1 serum, neutralization-test-positive, contained group B HI antibodies-namely, to yellow fever, West Nile, and Wesselsbron viruses. Sera collected from 7 patas monkeys (Erythro- cebus patas) at Bassa, near Jos, a year after the 1969 epidemic, were negative in yellow fever neutraliza- tion tests. DISCUSSION The yellow fever outbreak in the Jos region in 1969 was the first recognized occurrence of epidemic yellow fever in Nigeria in 17 years. Previous epi- demics occurred in Ogbomosho in 1946 and in Eastern Nigeria in 1951-53 (MacNamara, 1954). Although many hospitalized cases in the 1969 epi- demic presented a classical textbook picture of yellow fever, physicians were initially reluctant to diagnose yellow fever and instead applied terms such as " nephrotoxic hepatitis " and " plateau jaundice " to the disease. The latter term was, in fact, a well- remembered diagnosis, having been employed dur- ing similar epidemics on the plateau in 1931 and in 1951-52. Once the true nature of the illness was suspected, laboratory diagnosis was relatively easy. Virus could readily be isolated from acute-phase serum, and complement-fixing antibody responses pointed in general towards infection with yellow fever virus. On one occasion, virus was recovered from blood serum collected in the field and kept at the ambient temperature for 24 hours during transport to the laboratory. Another virus-positive serum was kept for 6 days in the freezing compartment of a refri- gerator. Virus in serum samples therefore seems reasonably stable, and isolation may be accom- plished under less than ideal circumstances. Although the pathological diagnosis of yellow fever was straightforward in most cases, liver speci- mens from several patients with excellent clinical and serological bases for a diagnosis of yellow fever showed nonspecific pathological changes (Francis et al., 1972). The estimate of 252 hospitalized cases of yellow fever is conservative since no admis- sions prior to 7 September are included. Hepatitis is believed to have continued at its usual level of activity throughout the yellow fever epi- demic. At Evangel Hospital, where reliable records were available, an average of 5 patients with hepa- titis were admitted between 7 September and the end of the year each year from 1964 to 1968. In 1969, during the same period, hepatitis was diagnosed in only 5 patients while 27 diagnoses of yellow fever were made. No data are available for the overall case-fatality ratio. The value of 38% for hospitalized cases is a minimum since patients were frequently removed from hospital by a relative when it appeared that death was inevitable. An attempt was made to carry out a mortality survey in one district with the help of lay personnel, but spot checks by different inspectors revealed wide discrepancies and the data had to be discounted. Surveys of this nature are rendered extremely difficult by cultural inhibi- tions with respect to references to deceased persons. In at least one population, the students and faculty at Gindiri Teacher Training College, where 17% of those sampled showed evidence of having recently been infected with yellow fever, there were no deaths. On the other hand, reports of multiple deaths in a single household were common. The marked preponderance of hospitalized males caused considerable speculation, the male: female 649 D. E. CAREY ET AL. ratio being 4: 1. This ratio may simply reflect the greater likelihood that males will seek hospitaliza- tion. A less likely possibility is that the disease is more severe in males. In neither the Du nor the Gindiri survey were significant differences found between the sexes with respect to the prevalence of yellow fever complement-fixing antibody, an indi- cation that in those populations, at least, there was no evidence of higher infection rates among males. At Gindiri, furthermore, equal proportions of males and females who were positive for complement fixation reported having experienced a febrile illness during the 2 preceding months. Thus mild clinical infection appeared to be equally likely in males and females. Complement-fixing antibody for yellow fever virus detected in sera collected in November would have been evoked by infection during the peak epidemic period 2 weeks before, or earlier. The values of 7% and 17% found, respectively, for the village of Du and a residential school therefore represent low esti- mates of the actual proportion of each population infected during the whole of the epidemic. Approxi- mately 40% of all hospitalized cases, for instance, were admitted after the surveys were made. A truer estimate for community involvement might be closer to 10% of the approximately 1 million people living outside Jos on the plateau and in the immediately adjacent areas. The epidemic reached a peak at a time when the rainfall was rapidly declining and temperatures were falling. Indeed, the drying-up of potential habitats for mosquito larvae in November greatly hampered the search for a vector (Lee & Moore, 1972). It is believed that these climatic factors were principally responsible for the rapid termination of the epidemic by December 1969. From then until the end of December 1970, no other cases of yellow fever were detected on the plateau. The establishment of a massive programme of immunization, in which approximately 20% of the plateau population were given 17D vaccine, may have served as a barrier to the reintroduction of yellow fever into the human population in 1970. Jos, the one large urban centre on the plateau, was spared during the 1969 outbreak. Since the town population was presumably just as susceptible as the rural population to yellow fever, the small number of cases can probably be explained by the absence of a vector in the town. Aedes aegypti mos- quitos were not found in Jos but this species is present in both Lagos and Ibadan, and the failure to detect yellow fever in these areas remains unex- plained. It has been suggested that the low incidence of clinical yellow fever in certain parts of Nigeria may be due to " tolerance . . . induced by prior infections with related virus " (MacNamara et al., 1959). Although the vector mosquito was not definitely identified, it is believed to have been A. luteocephalus. One isolation of yellow fever virus was made from a mixed collection of Stegomyia spp. (Lee & Moore, 1972). Whether yellow fever virus existed as an endemic focus on the plateau or was introduced from another area is not known but the latter is suspected to be the case. Inquiries revealed the probable occurrence of an outbreak of yellow fever in Benue Province, south of Plateau Province, in 1968. Dissemination of virus to the north by infected persons could then have initiated the epidemic among a largely suscept- ible population at a time when other ecological fac- tors were particularly favourable. Historical evid- ence suggests that this process may take place every 20-30 years on the plateau. Neutralization tests with nonhuman primate sera from other areas of Nigeria indicated that bush- babies and monkeys (including baboons) could be involved in a yellow fever cycle in Nigeria. HI tests, however, led to questions being raised about the specificity of the neutralization test results. Neutral- izing antibody has been reported previously in West African monkeys (Findlay & MacCallum, 1937), but such tests must be interpreted with caution in view of the reactions that occur also with closely related agents (MacNamara et al., 1959). The nature and importance, if any, of an extra-human reservoir of yellow fever in West Africa remains to be determined. ACKNOWLEDGEMENTS The authors are grateful to officials of the Federal Ministry of Health, Lagos, Nigeria, as well as to the various state Ministries for facilitating these investiga- tions. Mr H. A. Odelola, Mr E. O'Connor, and Mr D. C. Wilson performed most of the serological work. Dr 0. R. Causey and Dr D. L. Moore were responsible for the virus isolations. Dr T. I. Francis submitted a number of the specimens for diagnosis. 650 EPIDEMIOLOGY OF A YELLOW FEVER EPIDEMIC IN NIGERIA 651 RESUME ASPECTS JEPIDEMIOLOGIQUES DE L'EPIDEMIE DE FIEVRE JAUNE DE 1969 AU NIGERIA Un certain nombre d'investigations ont e menees durant 1'epidemie de fievre jaune qui a sevi en 1969 sur le plateau de Jos au Nigeria. On a obtenu la confirmation du diagnostic au laboratoire pour 55 malades: dans 15 cas, par l'isolement du virus; dans 25 cas, par les examens sero- logiques; dans 12 cas, par l'etude des biopsies hepatiques; dans 3 cas, par 1'emploi de plusieurs de ces methodes. L'epidemie est devenue manifeste en septembre, mais il se peut que des cas soient survenus en juillet dej"a. Aucun cas n'a e enregistre apres le mois de decembre. L'examen des dossiers, dans sept hopitaux de la region, a reve1l qu'au total 307 malades y avaient ete admis entre le 7 septembre et le 31 decembre avec le diagnostic de fievre jaune, d'hepatite ou d'ictere. Parmi eux, 116 sont decedes. La plupart des patients hospitalises etaient de jeunes adultes, la proportion des sujets de sexe masculin etant de 4 pour un sujet de sexe feminin. En novembre, des examens serologiques ont ete prati- ques chez des habitants des villages de Gindiri et Du. Dans le premier d'entre eux, 17 %, et dans le second, 7% des serums renfermaient des anticorps fixant le compl6- ment pour le virus amaril. La fievre jaune a sevi particulierement en milieu rural. Le fait que la population urbaine ait ete en grande partie epargnee peut s'expliquer par l'absence d'Aedes aegypti. Bien qu'on n'ait pas identifid avec certitude le vecteur de l'epidemie, il semble qu'on puisse incriminer Ae. luteo- cephalus. REFERENCES Causey, 0. R. & Kemp, G. E. (1968) Niger. J. Sci., 2, 131-135 Findlay, G. M. & MacCallum, F. 0. (1937) Trans. roy Soc. trop. Med. Hyg., 31, 103-106 Francis, T. I. et al. (1972) Bull. Wld Hlth Org., 46, 659 Lee, V. H. (1972) Bull. Wld Hlth Org., 46, 641 Lee, V. H. & Moore, D. L. (1972) Bull. Wld Hlth Org., 46, 669 MacNamara, F. N. (1954) Bull. Wld Hlth Org., 11, 391401 MacNamara, F. et al. (1959) Trans. roy. Soc. trop. Med. Hyg., 53, 202-212 Weinbren, M. P. (1958) East African Virus Research Institute, Report No. 8, p. 38 7