Previous PageIndexNext Page

AFRICAN HAEMORRHAGIC FEVER IN THE SOUTHERN SUDAN, 1976: THE CLINICAL MANIFESTATIONS

D.H. SMITH (1), D.P. FRANCIS (2), D.I.H. SIMPSON (3)

1. Tana Project, Ministry of Health, P.O.Box 53131, Nairobi, Kenya.
2. Center for Disease Control, Atlanta, U.S.A.
3. London School of Hygiene and Tropical Medicine, England.

The clinical description of African Haemorrhagic Fever in the Sudan was obtained from a variety of sources. In Maridi, clinical details of patients were obtained both from active cases, and hospital records and by interviewing recovered cases or close family members. In Nzara, the descriptions were based entirely on retrospective interviews.

Of the total 280 cases occurring in the area, adequate clinical descriptions were obtained from 183. Information was collected on a questionnaire form designed prior to the investigation utilizing the symptomatology described in previous outbreaks of Marburg virus disease.

For the purpose of this paper a case has been defined as at least 2 days fever plus either gastrointestinal symptoms, chest pain or haemorrhagic remarkations and a clear history of contact. In Nzara where apparent primary cases were occurring, the clinical definition demanded haemorrhagic features.

The Onset of the Disease

Initially the disease presented as a progressive febrile, 'flu-like' illness. Earliest complaints included a fever, usually between 100 and 102ºF, severe headache and generalised myalgia in all cases. The course of the illness was steadily progressive. Within a few days patients became progressively more ill, exhibiting a uniform appearance with deeply sunken eyes and a fixed expressionface described as "mask like" or "ghost like". During the early stages of illness, patients, especially in Nzara, presented at the hospital, where were given injections of chloroquine and antibiotics as outpatients.

Other early symptoms included a dryness, or soreness of the throat which occured in 63% of all cases. This appeared to make the individual disinclined at or drink but was only rarely described as painful. Chest pain was an other early and frequent symptom occurring in 83% of cases, sometimes severe and commonly described in the lower costal areas occasionally clearly pleuritic and associated with a dry cough.

These initial symptoms became progressively more severe over the first four or five days, by which time most patients had presented themselves to the hospital. They appeared toxic, complained of severe headache and myalgia.

The myalgia often made patients reluctant to be examined. Muscles were tender to palpation, joint movements were painful and many patients complained especially of lower back and lumbo-sacral pain.

Gastrointestinal Symptoms

Gastrointestinal symptoms were a prominent feature of this outbreak.

Usually starting towards the end of the first week but occasionally as early as the second or third day patients commonly developed gastrointestinal symptoms. The most frequent was diarrhoea which occurred in 81% of cases. It began abruptly and lasted for about seven days in those who survived. Vomiting occurred less frequently (in 59%), starting after the onset of diarrhoea.

Accompanying the diarrhoea and vomiting many patients described colicy abdominal pain and developed a profound anorexia.

Cutaneous Manifestations

In Maridi, a number of acute cases were observed to develop a rash around the 5th to 7th day and we consider that most cases do exhibit cutaneous manifestations. Despite the large number of retrospective interviews in our patients, 52% reported either a noticeable rash during the illness or subsequent desquamation. The rash, was rather measles like, papular or maculo-papular and predominantly seen on the upper arms, flexor surfaces of the forearms and upper legs. Desquamation when it occurred, took place some 10 to 14 days after onset of disease and appeared in the same sites but also especially on the palms of the hands and soles of the feet.

Haemorrhagic Features

Haemorrhagic manifestations were both a characteristic feature and a prognostic indicator in Maridi and Nzara. Virtually all of the fatal cases had visible blood loss (91%), whilst 71% of all documented cases had haemorrhagic features.

The most frequent and often severe manifestation was gastrointestinal haemorrhage, which occurred in 59% of cases and 86% of fatal cases. This took the form either of watery diarrhoea with fresh blood, malaena stools, or vomiting of fresh blood.

Whilst gastrointestinal haemorrhage was the most common expression of the haemorrhagic diathesis, bleeding was noted in a variety of other situations. Bleeding from the nose, mouth and gums was the second most frequent, occurring in 50% of fatal cases in Nzara and was often severe. Subconjunctival haemorrhages were commonly described. Vaginal haemorrhage was reported and a few patient had cutaneous haemorrhages. Haematuria appeared unusual.

Haemorrhagic features occurred after about the 5th day of illness and reached a maximum on the 10th day. The severity of the disease appeared to be directly related both to the extent and the severity of the haemorrhagic symptoms.

Central Nervous System

Symptoms referable to the central nervous system appeared to be frequent in this outbreak. Neck stiffness was reported especially in the more severely ill Spinal fluid obtained in these early cases was reportedly clear macroscopically.

Many patients exhibited bizzare behaviour - patients tended to abscond from hospital and behaved in an inappropriate manner, stripping off their clothes and wandering about the hospital in a confused state. One patient developed a terminal hemiplegia whilst a further patient was readmitted following recovery with overtly psychotic symptoms.

Physical Examination

Physical examination presented few characteristic features. The most characteristics was the general appearance of the patient, even in the early stages of the disease. The drawn, mask-like features, sunken eyes and loss of skin turgor came to be recognized as characteristic often supplemented with dry mucous membranes and oral fissuring. Patients both resisted and resented physical examination.

The posterior pharyngeal wall was injected. Neck stiffness was demonstrable in several of the more severely ill individuals.

The abdomen was soft and neither the liver nor spleen were palpable although tenderness was observed in the epigastrium and below the right subcostal margin. Jaundice was not observed.

The Course of the Illness

Severe cases demonstrated a relentless deterioration. Death most commonly Occurred on the 9th day although ranging from 2 days to 21 days after the onset The majority of deaths occurred predictably in severely ill patients usually exhibiting most of the classical features of the disease with severe haemorrhagic features. However, death also occurred in individuals convalescing from the infection. These deaths were sudden and unexpected. Recovered cases suffered a prolonged convalescence, often with continuing headache and profound lethargy, continuing for up to several months after the acute illness.

Mortality

The overall mortality in the Sudan outbreak was 51%. This mortality was deduced from the total number of cases considered to be due to AHF on clinical and epidemiological grounds. In Maridi where a higher proportion of clinical cases had serological evidence of infection the overall mortality was 54%.

In Maridi there was little evidence to indicate any alteration in mortality as the outbreak progressed. In Nzara, however the mortality may have been higher at the start of the outbreak falling from 88% in July to 62% in August and 38% in September (Table 1). However the numbers, especially in July, are small.

TABLE I
MORTALITY BY MONTH OF INFECTION IN NZARA



July


August


September


October


Deaths


7


13


14


0


Total


8


21


37


4


Per Cent


88


62


38


-

Comparison of Maridi and Nzara

The absence of serological confirmation in the majority of surviving patients in Nzara has lead us to compare the clinical features of cases in the two outbreaks (Table 2). Both the time of onset of the various clinical features and their prevalence suggest that the two groups were consistent. The only evident differences were the frequency of chest pain, vomiting and cutaneous manifestations, all of which could be explained by the retrospective nature of the Nzara surveys. Haemorrhagic manifestations, severe symptomatology and mortality were consistent in the two groups.

TABLE 2
CLINICAL SYMPTOMS - MARIDI AND NZARA, SUDAN, 1976


Symptoms


Frequency (183 cases)


 

Maridi


Nzara


Fever


100%


100


100


Headache


100%


100


96


Chest Pain


83%


87


76


Diarrhoea


81%


84


76


Vomiting


59%


70


43


Dry Painful Throat


63%


63


62


Rash or Desquamation


52%


64


35


Cough


49%


60


33


Bleeding (any)


71%

   

Melena


59%

   

Bleeding (recovered cases)


48%

   

Bleeding (fatal cases)


91%

   

Asymptomatic Cases

The available evidence both in Nzara and Maridi, suggests that very mild or asymptomatic cases occurred. The sero-positive cases detected in the Nzara Cotton Factory and subsequently followed up indicate that half had experienced no illness over the past year whilst the others had had relatively mild febrile illnesses. Similarly, in Maridi, close contacts of cases were found to have serological evidence of infection but no history of recognizable illness.


DISCUSSION

The differential diagnosis of haemorrhagic fever especially in individual patients presents considerable difficulties in rural African populations. In an individual patient, the clinician must consider bacterial causes such as meningococcal septicaemia, septicaemic plague and relapsing fever: protozoal causes including malaria and trypanosomiasis as well as a variety of viral infections of which yellow fever, Marburg virus disease, Lassa fever and now African haemorrhagic fever are the more important.

However, when a cluster of cases occurs with a prodromal febrile illness followed by a haemorrhagic diathesis in a high proportion of cases and when transmission from person to person is observed amongst close contacts of cases especially when involving hospital staff, the possible aetiological candidates are considerably reduced. In Africa, Lassa fever, Marburg virus disease and now African Haemorrhagic Fever appear the most likely causes.

These three diseases present many features in common. The prodromal illness, vomiting, chest pain and rash are almost identical in Lassa and AHF. The pharyngitis in Lassa is commonly pronounced and the conjunctivitis severe and associated with periorbital swelling in contrast to AHF, where both the pharyngitis and conjunctivitis are rarely severe. Diarrhoea also occurs in both infections, although of greater severity in AHF.

The difficulties in clinical differentiation between AHF and Lassa are even greater when AHF is compared to Marburg virus disease. The symptomatology of the Sudan outbreak is substantially the same as that described in the two outbreaks of Marburg except for the frequency of chest pain which was rare in the previous outbreaks of Marburg. However chest pain was also uncommon in the Zaire outbreak of AHF.

The laboratory facilities available in Nzara and Maridi, or for that matter any similar hospital in rural Africa, provide little assistance in the diagnosis apart from the exclusion of alternative aetiologies. Few laboratory investigations were carried out in the Sudan outbreak but it is not anticipated that the findings would differ from the detailed studies carried out in the two previous outbreaks of Marburg.

The clinician practising in the rural hospital therefore has to rely on the clinical and epidemiological information available. Suspicion would then set in motion a train of events including the collection of appropriate samples for sophisticated virological investigation, the possible use of disease specific immune sera based perhaps more on an epidemiological assessment than the clinical features and the institution of personal, institutional and community measures to contain transmission.

The mobilization of adequate logistics to permit the collection, transport and processing of sophisticated virological samples presents numerous difficulties in remote areas of rural Africa. The subsequent low isolation rate obtained from Sudan material and the apparent transience of detectable antibodies further complicate an adequate investigation of such outbreaks which by their very nature tend to arise in remote-areas with limited resources. These technical and logistic questions remain of paramount importance to the physician practising in remote rural populations as well as to those responsible for epidemiological surveillance and public health in countries at risk.

Previous PageTop Of PageNext Page