10 Fever in Tanzania

John Crump 30/08/2018

What are challenges for febrile in the tropics?

Malaria vs non-malaria

10.1 Clinical Problem (in 2002)

Primarily human infections, plus zoonoses.

Fever is common reason for admission, but the lab was limited. So most febrile were treated empicivally for malaria. But poor response clinically, and a high case fatality risk.

So why were they having a poor response to antimalarials, why were they dying?

In any 2 week period, about a 1/4 of children will have a fever

This provides a diagnostic challenge. In the abscence of a lot of diagnostic services, or epidemiological info, its difficult to know what the cause is, or what the best treatment would be.

So in this setting there is a lot of syndrome based clinical management. One set of guides would be integrated managment of childhood, or adult illnesses. Provided by the WHO.

You would follow an algorithm of a child w cough, fever, sob -> Pneumonia -> Abx

But these algorithms don’t really work for undifferentiated fever w no clinical signs.

2006 Malaria guidelines state: The signs and symptoms of malaria are non specific. So you may need to assume that it’s malaria when you’re in a malaria endemic area. (But this only works when prevalence is high!)

Malaria incidence is falling globally though, thanks to insecticides, drug treatment, and surveillance. Then later w/ indoor residual spraying, insecticide-treated nets, rapid diagnostic tests and artemesinin based therapies. These later strategies have resulted in a fall globally in malaria cases and malaria deaths.

Plasmodium falciparum prevalence in people in the community in Africa, there is a fall from 2000, vs 2015. There is great regional variability.

So in Tanzania in 2004, those who met the clinical criteria for severe malaria, only about half were actually positive for malaria. So then the case fatality rate in the group who were slide negative were more likely to die (7% risk of deaths in slide positive cases vs 12% in slide negative)

So the WHO recognised that 2006 guidelines result in over treatment of malaria, and the 2010 guidelines now talk about needing prompt parasitological confirmation in patients before treatment. You should only treat clinically when tests are not available.

This was good for recognising when sick patients don’t have malaria, but what do they have?

So when RDT are available, you now use a lot more antibiotics. This is fine in sick cases, but what about milder outpatient cases, did they really need antibiotics?

In 2010 in Tanzania, you can see southern tanzania has high malaria prevalence, but northern places like moshi have v low prevalences. In KCMC, 61% of febrile inpatients were diagnosed clinically with malaria. About 1/3 were diagnosed with other infections. But when you compare this to lab testing, only 1.6 were caused by malaria. Including blood stream infections, a wide range of loads of different infections (See CRUMP JA, PLOS NEglect Trop Dis 2013; 7: e2324) Loads of these are zoonoses. These illnesses can cause growth failure, misscarriage, and milk drop in animals, affecting peoples livelihood.

Blood stream infections, you see a lot of cryptococcal disease and mycobacterial disease with HIV positive patients. Theres a lot of salmonella typhi in moshi also.

10.2 Resources on Causes of Fever GLobally

Non malarial febrile illness map

Etiology of Severe Febrile Illness. There are substantial heterogeneity in causes of fever over place and time. (Prasad N PLOS ONE 2015; 10: e0127962 )

10.3 Ongoing Research

10.3.1 Bacterial Zoonoses

10.3.1.1 Leptospirosis

Zoonotic pathogen, with multi-host epidemiology. Resevoir is wild animals (rodants and live stock), can shed parasite in urine for rest of their life. Some leptospirosis causes no disease in animals, depends on serovar and host. Humans can get infected by animals, or through water.

Humans usually have febrile illness. Loads of human disease happens in the “immune phase”. Humans may shed leptospira for short period, but not the resevoir that animals are.

Neglected zoonosis. Laboratory diagnosis is super complex. Cultures not useful in most phases, neither is PCR. As a result, it’s often missed out of burden of disease estimate, including the Institute for Health Metrics and Evaluation.

Leptospirosis is not just an urban disease, it happens in the tropics. Most data is lacking about resevoirs in africa. Acute human leptospirosis is data lacking in humans. Review around three years ago, showed existemce in africa.

SO to estimate leptospirosis. You can use sentinel case surveillence. There is a fair amount of variation in a year.

2 serogroups predominate in Northern Tanzania. Mini and Australis. Cattle, hedgehogs, dogs, sheep, pigs, donkeys, were all animals implicated. Recent survey showed people who worked with cattle waste, or grew rice in a paddy field.

Cattle had a 7.3% prevalence, kidneys/urine PCR leptospirosis infection.

There are some leptospirosis groups found only in humans. But there are 6 groups found in human and cattle. Both are serogroups whick can affect humans and cattle.

The further out into pasturalist areas, there’s a particular increased amount of leptospirosis in cattle. Their difference is shared animals and potential outcome.

10.4 Febrile Deaths

This historically data was focussing of severe disease. Two thirds of of world popuation live in areas without systems to accurately record deaths. Talking to family members after death, unlikely to get true cause of deathd.

Verbal autopsy as a method may be quite inccurate. Autopsy studies just generally work on subgroups.

And remember a big contributor to febrile deaths could and will be systems failings.