Acute bacterial infections and HIV disease
Charles F Gilks
Liverpool School of Tropical Medicine, Liverpool L3 SQA, UK
Some acute bacterial infections, notably those caused by the pneumococcus and
the non-typhi salmonellae, have not traditionally been considered as leading
diseases in tropical medicine, despite their ubiquitous distribution and impact on
health. The HIV/AIDS epidemic is forcing a re-evaluation of this position because
of their importance in immunosuppressed adults, particularly where exposure is
high and treatment relatively inadequate. The problem of acute bacterial disease
in HIV/AIDS is outlined in industrialised countries and contrasted with the
problem in tropical countries. Specific insights into HIV-related pneumococcal
disease and non-typhi salmonellosis that have come from work in the tropics are
then discussed. These infections need now to be recognised as an important
element of tropical medicine.
Correspondence to
Dr Charles F Gilks.
Liverpool School of
Tropical Medicine,
Pembroke Place,
Liverpool 13 5QA, UK
Acute bacterial infections have always been important causes of
morbidity and mortality in the tropics, in much the same way as they
were in temperate regions a century ago. In industrialised countries
nowadays, as a result of improvements in socio-economic status, widely
accessible good-quality medical services and highly effective antimicrobial therapy, pneumococcal pneumonia is clearly not the problem
it was1; typhoid and cholera2 are no longer endemic diseases; and
rheumatic fever is very rarely diagnosed3. The epidemiological transition
has been slow in many regions of the tropics, and acute bacterial
infections remain leading killers today4.
Despite their undoubted importance, several of the most common
acute bacterial infections have remained on the fringes of classical
tropical medicine, at least as diseases of adults. In particular, Grampositive cocci hardly attract any attention in standard UK5 or US6 text
books; and salmonellosis essentially means typhoid fever. While the
founders of tropical medicine would have been surprised to have to
consider the pneumococcus or the non-typhi Salmonellae as particular
tropical problems - infection by these pathogens was ubiquitous and
100 years ago pneumonia or food poisoning was very similar in London
or Lagos - today such an approach seems less valid. There are clearly
special features of these pathogens in the tropical regions, and it is
predictable that bacterial infections will receive more prominence in the
next 100 years of tropical medicine. One of the most important features
British Medical Bulletin 1998,54 (No. 2): 383-393
O The British Council 1998
Tropical medicine: achievements and prospects
is the rapid spread of HTV across resource-poor countries7, and the great
importance of acute bacterial infections in those immunosuppressed by
this virus.
The problem in industrialised countries
Early in the HTV epidemic in the US, it was recognised that there was an
increased susceptibility to some, but not all, bacterial and mycobacterial
infections, as well as to a restricted range of opportunistic parasitic and
fungal infections. In the first surveillance definition, recurrent non-typhi
Salmonella (NTS) bacteraemia was included as an AIDS-defining event
because it was such an unusual clinical problem and was highly
associated with, and predictive of, underlying HTV infection8. There was
also a much higher rate of bacterial pneumonia in drug users in excess
of their already higher susceptibility to pneumonia, and in homosexual
men. Most often the pneumococcus was implicated, but Haentophilus
influenzae and Staphlococcus aureus were also recovered9'10.
It was soon shown that HTV-infected patients had defective B-cell,
monocyte/macrophage and neutrophil function11"13 although it remains
largely unclear why HTV infection predisposes to a relatively restricted
range of pathogenic bacteria. It is debatable whether these virulent
bacteria and Mycobactium tuberculosis, all pathogens capable of
causing severe disease and death in Lmmunocompetent individuals,
should be considered as opportunistic pathogens merely because they
cause infection in a patient with HIV/AIDS. It is perhaps more useful,
and logical, nowadays to consider most acute HTV-related bacterial
pathogens as high-grade infections, particularly because they may occur
early on in the natural history of HTV infection; and reserve the term
opportunistic infection for those highly unusual problems characteristic
of marked immunosuppression and AIDS itself.
Pneumococcal pneumonia is now recognised as a leading early HTVrelated problem14; susceptibility increases with progressive immunosuppression15; a wider spectrum of disease with some unusual
manifestations is recognised16; and rates of pneumococcal bacteraemia
are estimated to be 100-fold higher in patients with AIDS17. There can
be a good response to therapy and survival from bacteraemic disease is
better than in some groups of more elderly patients with other risk
factors for pneumococcal disease14. As with many HFV-related
pathogens, there is a greatly elevated risk of subsequent infection
following the initial event, although the risks have not been well defined.
Recently, recurrent bacterial pneumonia, within 12 months of the initial
event, has been included as an AIDS defining event18. Since 1988, it has
384
British Medical Bulletin 1998;54 (No. 2)
Acute bacterial infections and HIV disease
been recommended that any person with HIV infection, irrespective of
the stage of disease, be offered pneumococcal polysaccharide vaccine.
These guidelines are based on the premise that the vaccine may be
immunogenic and is relatively safe and cheap; no efficacy trials have
been carried out19.
NTS bacteraemia is relatively rare, probably because of limited
environmental exposure in communities that have good public health
and hygiene. In one multi-centre study in the US, NTS bacteraemia was
ranked 20th out of 21 problems, and accounted for just 0.4% of major
clinical events20. Other Enterobacteriaceae, especially Escherichia coli,
but also Klebsiella spp, are well recognised in AIDS patients with acute
community-acquired bacteraemia21. H. influenzae infection, both with b
and other capsular types, is described particularly in patients with
marked immunosuppression22. The conjugate vaccine is immunogenic
but probably not of particular use because overall invasive disease is a
relatively infrequent event23. Pseudomonas septicaemia is also seen with
increased frequency with advanced immunosuppression and is
associated with central venous catheters; it may be either nosocomial or
community-acquired24. Staphylococcal infections are common especially
in patients with long lines and indwelling catheters and are seen with
higher frequency in intravenous drug users25.
There have been numerous case reports of unusual bacterial, or other,
pathogens causing acute or chronic septicaemic illnesses26, but few
strong associations have emerged. Rhodococcus equi is a cause of
unusual cavitary pulmonary lesions but tends to present chronically27.
There are no convincing data to suggest that organisms like Brucella or
Listeria are significantly associated with HIV, despite their intracellular
location.
The problem in tropical countries
Acute high-grade bacterial infections are, in general, far more common
than opportunistic infections in patients infected with HTV who live in
Africa. This is most clearly seen in cross-sectional studies of HTVinfected patients admitted to hospital. In two separate studies in
Nairobi, done 3 years apart with the same methodology, one-quarter of
patients were bacteraemic on initial assessment which included a single
5-10 ml blood culture28-29. In this population prior, self-administered
antimicrobial therapy was common and relatively indiscriminate.
Streptococcus pneumoniae and Salmonella typhimurium predominated.
Similar results were obtained in cross-sectional blood culture studies of
patients at or around the time of admission from central and western
British Medical Bulletin 1998;54 (No. 2)
385
Tropical medicine: achievements and prospects
Table 1 Bacteraemia and fungaemia in HIV-positive adults admitted t o hospital in Africa: a summary of crosssectional studies from East, Central and West Africa
Total with bacteraemia (%)
Non-typhi Salmonellae (%)
Strep, pneumoniae (%)
Mortality in patients bacteraemia (%)
C neoformans (%)
Nairobi, 1988/89
Kigali, 1989
(n = 63)*
Abidjan, 1991
(n = 292)31
26.3
10.5
7.4
58
1.1
23.3
6.1
7.4
NA
8.0**
15.3
7.9
1.5*
63
2.5
Nairobi, 1992
(n = 144)*
24.4
10.4
10.4
37
0.7
Abidjan, 1995
(no 199)*
19.6
11.6
2.0*
47
0.0
NA, data not available; 'significantly lower than other studies (P< 0.001); "significantly higher than other studies (P < 0.001).
Africa (Table I)30"32. The low rate of pneumococcal isolation in both
studies in Abidjan is explained by the patient population selected for
study. They were patients screened and referred to an Infectious Diseases
and AIDS Unit, whereas in the other studies, unscreened general medical
patients were studied.
In hospital, most pneumococci in blood are recovered from patients
with relatively typical lobar pneumonia. Patients with NTS bacteraemia
tend to present with a typhoid or enteric fever-like illness. There is some
geographical variability in the types of NTS recovered. Salmonella
enteritidis made up about 20% of NTS isolated from blood in West
Africa but was less than 5% in East and Central Africa. Such regional
differences have not yet been explained. It is interesting that there is
much more variation in cryptococcal fungaemia rates than in
bacteraemia rates. Rates are far higher in Central than East or West
Africa. Limited cohort data from East Africa confirm the high rates of
invasive pneumococcal disease and NTS bacteraemia33, that rise with
worsening immunosuppression34. So far, there are very few data from
other regions in the tropics, although it is probable that many of the
features seen in Africa will be common in other resource-poor
communities in Asia and the Americas.
The WHO has proposed a clinical staging system for HTV infection
and disease, to be used particularly in resource-poor countries where
laboratory indicators of stage or status, such as CD4 counts and viral
loads, are not widely available35. Because of their importance, acute
bacterial problems figure prominently in this clinical staging, especially
in the earlier stages: stage 2, indicative of mild disease, includes
recurrent upper respiratory infections {i.e. bacterial sinusitis); stage 3,
indicative of intermediate or moderate disease includes severe bacterial
infections (i.e. pneumonia, pyomyositis) as well as pulmonary
tuberculosis; and stage 4, equivalent to AIDS, includes NTS septicaemia.
In Nairobi rates of acute bacterial infection were high and more than
50% of patients on admission to hospital were in stage 2 or 3, or pre-AIDS;
386
British Medical Bulletin 1998;54 (No. 2)
Acute bacterial infections and HIV disease
they were generally not recognisable as patients with underlying HIV
infection. Similar figures have been reported from comprehensive
admission surveys in both Uganda36 and Zaire37. In contrast, the low CD4
counts noted in one of the Abidjan studies reflected the fact that the
patients had been referred to a specialist AIDS and infectious diseases
service32. One problem in interpreting CD4 counts in cross-sectional
studies is that they are reversibly suppressed in acute bacterial infection33.
An important consequence of this is that disease surveillance with any
clinical AIDS definition consistently misses at least half the morbidity,
and much of the mortality directly attributable to HTV; thus the true
impact of HIV/AIDS can be seriously underestimated38. Another
problem is that major shifts in hospital performance and care may not
be appreciated, in particular the crowding out and displacement of HTVuninfected patients by those with HTV-related disease39.
Although there are few data about the natural history of HTV7AIDS in
tropical as opposed to industrialised countries, it seems that the time for
progression to AIDS or stage 4 disease is similar, but that once AIDS
occurs, survival time is far shorter. Thus in Uganda, while progression
rates to AIDS are similar to those in developed countries, the subsequent
median survival time with AIDS was only 9.3 months40. In contrast,
Africans living in London have very similar disease progression and
survival to UK patients attending the same clinics41. Limited clinical and
post-mortem data about causes of death in HTV-infected patients in
Africa suggest that a different spectrum of pathogens is responsible; and
that acute high-grade bacterial (and mycobacterial) pathogens
predominate42.
There are a variety of interlinked environmental and socio-economic
reasons that explain the different features of HIV disease and AIDS in
the tropics, especially the clinical spectrum and early death. In
particular, this relates to the high background prevalence of acute
bacterial and mycobacterial disease. Pathogen exposure in poor,
overcrowded urban and rural settlements with inadequate sanitation
and limited clean water is far higher than in the more affluent and better
housed communities more typical of industrialised countries. In
children, this is reflected in very high rates of acute respiratory infections
and diarrhoeal disease that are several fold higher than in the US or
Europe, and tuberculosis is endemic; these syndromes are, therefore, an
important cause of severe paediatric morbidity and mortality. The
consequences for HTV-infected adults are hardly surprising: a far higher
rate of acute bacterial and mycobacterial infection presenting relatively
early in the course of progressive immunosuppression. A poor
nutritional status may further compromise an HTV-infected individual's
resistance to acute bacterial and mycobacterial disease, although so far
there are no firm data supporting this.
British Medical Bulletin 1998;54 (No. 2)
387
Tropical medicine: achievements and prospects
Once sick, HIV-positive individuals tend to fare badly. Appropriate
health seeking behaviour may be adversely affected by poverty. Far more
patients are forced to present relatively late in the course of their acute
intercurrent illness. The standard of health-care available may be suboptimal: health workers may not be always available, well motivated or
adequately supervised; facilities are often overcrowded and underfunded; and even essential drugs may be scarce or unavailable. Under
these circumstances, survival with any severe and life-threatening disease
is likely to be greatly compromised by the inadequate level of health care
to which the individual with HTV infection has access. Many poor, sick
individuals die with their first or second episode of intercurrent disease.
Neither antiretroviral therapy nor specific disease prophylaxis are
widely used or generally recommended.
Insights into pneumococcal disease
Several features of the interaction of HTV with Strep, pneumoniae have
been highlighted by studies done in the tropics, but it is still not clear
exactly why individuals with a specific loss of CD4 lymphocytes are so
susceptible to one of the classic B-cell pathogens. One factor may be
related to nasopharyngeal carriage and disrupted mucosal immunity. In
the US, it appears that nasopharyngeal carriage in adults is not affected
by HIV infection43, although it is in children44. Where exposure is
relatively intense, it seems that pneumococcal carriage rates are
approximately double in those with HTV infection. In Nairobi, rates in
adults ranged from 21-25%, about double those in HFV-uninfected
people (unpublished data).
In Uganda, the efficacy of pneumococcal vaccine in HTV/AIDS is
currently under study. This includes a detailed investigation of carriage,
in relation to different levels of immunosuppression, serotype distribution, extent of antimicrobial resistance, and seasonal patterns.
Preliminary data indicate a higher rate of carriage with HIV infection (N
French and C Gilks, unpublished data). At higher rates, intensities and
levels of carriage, more nasopharyngeal secretions may be infected with
the pneumococcus, thus predisposing to higher rates of disease by
inhalation and aspiration into the alveolar spaces or by direct extension
into the blood.
Very few cohort data have emerged from the US or Europe about HIV
and the risks of pneumococcal infection. In part this is because, in such
countries, bacterial pneumonia is rarely life-threatening and therapy is
usually started rapidly, often in the community and early on in the
disease episode. Prophylactic antibiotics, for whatever indication, will
388
British Medical Bulletin 1998;54 (No. 2)
Acute bacterial infections and HIV disease
reduce the rate of bacterial pneumonia45. In one cohort of female sex
workers in Kenya, an annual rate of invasive disease of 4.25% was
described, and the relative risk for invasive disease with HTV infection was
17.8 33 . This relative risk is comparable to that seen in TB disease with HTV
infection. In Uganda, rates of pneumococcal bacteraemia, unadjusted for
whether placebo or vaccine has been administered, show a sharp rise with
falling CD4 count at enrollment: rates of 10/1000 per person year have
been found34 in all study recruits with CD4 count at entry of over
500/mm 3 rising to 29/1000 per person years in people with CD4 counts
ranging between 200 and 500/mm3 and 52/1000 per person years in those
with CD4 counts below 200/mm3.
Mortality in patients with HTV-related pneumococcal disease can be
very low, even when there is associated bacteraemia14"33, especially if
disease is diagnosed early and appropriate therapy is started rapidly.
People with HTV tend to be younger and with less serious underlying
problems than many HTV-seronegative patients with pneumococcal
bacteraemia. When presentation is delayed and pulmonary disease is well
established, mortality may be more than twice that seen in patients with
pneumonia who are HTV-uninfected. Mortality rates in excess of 15%
have been seen in HTV-infected young adults presenting for hospital care,
compared to about 8% for HTV-uninfected. Both rates are far higher than
in a population of similar age in the US or Europe, where a 2 - 3 %
mortality rate would be expected. Inadequate resources for appropriate
health-seeking behaviour is clearly a major obstacle to effective HTV/AIDS
disease management in Africa.
Those who survive the acute episode are predisposed to recurrent
problems or problems with other HTV-related pathogens. This is wellrecognised in temperate countries, and indeed justified the inclusion of
recurrent bacterial pneumonia within 1 year as an AIDS-defining clinical
event. However, several issues are unresolved. Are relapses attributable to
the same organism failing to be cleared fully or persisting in the
nasopharynx, or are they re-infections? In Nairobi, more than 90% of
cases were clearly re-infected with serotypes quite different from the initial
infecting strain33. What is the average time to relapse? In Nairobi, this was
just under 6 months; most occurred within 1 year. What is the rate of reinfection? In Nairobi, it was 264/1000 person years; very similar rates are
being described in Uganda.
Penicillin resistant pneumococci have recently been 'discovered' by the
various initiatives for emerging diseases now heavily involved in
surveillance, both regionally and globally. In Nairobi, the background
level of penicillin resistance was of the order of 2 5 % . One very alarming
observation was association of penicillin resistance (and other drugresistant phenotypes) with underlying HTV infection46. This can probably
be explained by higher antimicrobial use in HIV-positive patients,
British Medical Bulletin 1998;54 (No. 2)
389
Tropical medicine: achievements and prospects
particularly with sub-therapeutic dosages or incomplete treatment courses.
With high rates of nasopharyngeal carriage, there exists the clear and
alarming potential for further, and perhaps more rapid, spread and
evolution of penicillin resistance. If this were to be the case, the consequences for the tropics could be truly disastrous.
Options for disease prophylaxis include either vaccination or long-term
cotrimoxazole, presumably for life. Both are being evaluated in Africa and
results should be available in late 1998. If vaccination works, the issue
will be cost; the manufacturer's price for the single dose vial is currently
about US$ 12. The vaccine itself is simple to use and needs only a single
injection; there are no adverse public health consequences of failure19. If
cotrimoxazole prophylaxis works, the operational issue is compliance, as
well as the accrued costs for drugs and for clinical follow-up. As with
isoniazid prophylaxis for TB therapy, the biggest constraint will be longterm adherence. The public health consequences of a failing prophylaxis
intervention are likely to be complex, but will almost certainly include
higher rates of cotrimoxazole resistance.
Insights into Gram-negative infections
With high exposure, several Enterobacteriaceae are common causes of
bacteraemia and septicaemia in adults with HTV infection in the tropics.
The most important are the Salmonellae, as they are in temperate zones.
In Africa, the non-typhi salmonellae (NTS), particularly 5. typhimurium
and to a lesser extent S. enteritidis, predominate. There is no association
of S. typhi or S. paratyphi with underlying infection28"33. Typhoid is a
virulent infection, whether the patient has adequate T-cell immunity or
not, whereas NTS are only capable of regularly causing an enteric feverlike illness with bacteraemia in immunosuppressed adults. Consistently,
about 1 in 10 adults admitted to hospital with underlying HTV infection
has NTS bacteraemia (Table 1), and mortality is high.
NTS may not predominate in other tropical regions. In South America,
one study has shown an association of HTV infection with S. paratyphi
infection47. Preliminary data from Vietnam (C. Parry, personal communication) suggest that while the NTSs are uncommon in HTV infection,
classical 5. typhi is associated with immunosuppression. If these intriguing
regional differences are confirmed, they suggest great differences in
salmonella epidemiology, important variations in virulence or susceptibility, or perhaps differences in first-line community therapy with
more powerful broad-spectrum antibiotics such as the quinolones.
Although NTS infections are highly associated with HTV infection
across Africa, little descriptive epidemiology and few clinical case series
390
British Medial Bulletin 1998;54 (No. 2)
Acute bacterial infections and HIV disease
have been published. Limited data from Nairobi showed an overall
bacteraemia rate in a cohort of female sex workers of 16/1000 person
years of observation33. In Uganda, NTS bacteraemia is rare in adults with
CD4 counts above 500, is seen at a rate of 5/1000 in adults with CD4
counts between 2-500, and occurs at a rate of about 75/1000 in adults
with CD4 counts below 20043. The commonest presentation is with an
enteric fever-like illness (which before the era of HTV would have been
called clinical typhoid). Blind therapy can be difficult because antibiotic
resistance rates in NTS across Africa are high but variable. First line
empirical therapy also depends on what is available and affordable. Ideally
a quinolone or third generation cephalosporin should be used; however,
chloramphenicol or ampicillin with gentamicin were effective therapy in
Kenya, when the more expensive drugs were not generally available.
Mortality can be very high particularly when there is little awareness of
the problem. When the problem was first described in Nairobi, mortality
was 80% but with better recognition and more appropriate therapy
mortality declined to 29%28>29. Under trial conditions, mortality was
reduced to 12% 33 . As with most other HTV-related infections, survivors
suffer high recurrence rates. In a limited study in Nairobi, the recurrence
rate was 26%. Both re-infection and true relapse (which responded well
to a resumption of the same treatment regimen) were seen in about equal
numbers (C.F. Gilks, unpublished data) 33 .
Other important Gram-negative infections include E. coli, Klebsiella
and Shigella spp. All appear to occur at much higher rates in HTV
immunosuppressed individuals, but no large case series have yet been
published. In some patients, there is an obvious association with bowel or
urinary tract infection. In Nairobi, no special phenotypes of E. coli were
responsible for bacteraemia or chronic diarrhoea. In particular, there was
no obvious excess of EAEC, EIEC or other recognised virulent phenotypes
(C.F. Gilks et al, unpublished data) 48 .
The only option for prevention is cotrimoxazole prophylaxis. Given the
generally high, but heterogeneous, rate of drug resistance, antimicrobial
prophylaxis may not be particularly effective. Long-term compliance may
be difficult. Another unresolved issue in therapy is whether long-term
secondary prophylaxis is needed after the initial episode of bacteraemia.
Other simpler strategies, such as identifying risky behaviour, or the likely
sources of NTS exposure, have not yet been explored. Vaccination against
typhoid may be an option in those areas where S. typhi infection is
genuinely associated with underlying HTV infection. In Africa, this will have
little benefit and, at present, there are no NTS vaccines for human use.
References
1
Heffron R. Pneumonia with Special Reference to Pneumococcus Lobar Pneumonia. New York,
NY: The Commonwealth Fund, 1939
British Medical Bulletin 1998;54 (No. 2)
391
Tropical medicine: achievements and prospects
2 Cox FEG. (ed) Illustrated History of Tropical Diseases. London: The Wellcome Trust, 1996
3 Stollerman GH. Rheumatic fever. Lancet 1997; 349: 935-42
4. Feachem RG, Jamison DT. (eds) Disease and Mortality in sub-Saharan Africa. Oxford: The
World Bank/Oxford University Press, 1991
5 Cook GC. (ed) Manson's Tropical Diseases, 20th edn. London: WB Saunders, 1996
6 Strickland GT. (ed) Hunter's Tropical Medicine, 7th edn. London: WB Saunders, 1991
7 World Health Organization. HIV/AIDS: the global epidemic. Wkly Epidemiol Rec 1997; 72: 17-24
8 Centers for Disease Control. Revision of the case definition of AIDS for national reporting United States. MMWR CDC Surveil Summ 1985; 34: 373-5
9 Stoneburner RL, Des Jarlais DC, Benezra D et al. A larger spectrum of severe HTV-1 related
disease in intravenous drug users in New York city. Science 1988; 242: 916-9
10 Witt DJ, Caven DE, McCabe WR. Bacterial infections in adult patients with AIDS and AIDSrelated complex. Am J Med 1987; 82: 900-6
11 Ammann AJ, Schiffman G, Abrams D et al. B-cell immunodeficiency in AIDS. JAMA 1984; 251:
1447-9
12 Poli G, Bottazzi B, Acero R et al. Monocyte function in intravenous drug abusers with AIDS:
selective impairment of chemotaxis. Clin Exp Immunol 1985; 62: 136—42
13 Lazzarin A, Uberti-Foppa C, Galli M et al. Impairment of polymorphonuclear leucocyte
function in patients with AIDS. Cltn Exp Immunol 1986; 65: 105-11
14 Janoff EN, Breiman RF, Daley CL, Hopewell PC. Pneumococcal disease during HTV infection.
Ann Intern Med 1992; 117: 314-24
15 Schuchat A, Broome CV, Hightowcr A, Costa SJ, Parkin W. Use of surveillance data for invasive
pneumococcal disease to estimate the size of the immunosuppressed HTV-infected population.
JAMA 1991; 265: 3275-9
16 Barradas MCR, Musher DM, Hamill RJ et al. Unusual manifestations of pneumococcal
infection m HTV-infected patients: the past revisited. Chn Infect Dis 1992; 14: 192-9
17 Redd SC, Rutherford GW, Sande M et al. The role of HTV infection in pneumococcal
bacteraemia in San Fransisco residents./ Infect Dis 1990; 162: 1012-7
18 Centers for Disease Control. 1993 revised classification system for HTV infection and expanded
case definition for AIDS amongst adolescents and adults. MMWR CDC Surveil Summ 1992;
41 (R17); 1-19
19 Gilks CF. HTV and pneumococcal infection in Africa; clinical, epidemiological and preventative
aspects. Trans R Soc Trop Med Hyg 1997; 91: 627-31
20 Chan ISF, Neaton JD, Saravolatz LD et al. Frequencies of opportunistic diseases prior to death
among HIV-infected persons. AIDS 1995; 9: 1145-51
21 Krumholz HM, Sande MA, Lo B. Community-acquired bacteraemia in patients with AIDS:
clinical presentation, bacteriology and outcome. Am J Med 1989; 86: 776-9
22 Steinhart R, Reingold AL, Taylor F, Anderson G, Wenger JD. Invasive Haemophilus influenzae
infections in men with HIV infection. JAMA 1992; 268: 3350-2
23 Steinhoff MC, Auerbacj BS, Nelson KE et al. Antibody responses to Haemophilus influenzae
type B vaccines in men with HIV infection. N Engl J Med 1991; 325: 1837-42
24 Nelson MR, Shanson DC, Barter GJ, Hawkins DA, Gazzard BG. Pseudomonas septicaemia
associated with HIV. AIDS 1991; 5: 761-3
25 Jacobson MA, Gellermann H, Chambers H. Staphylococcus aureus bacteraemia and recurrent
staphylococcal infection in patients with AIDS and AIDS-related complex. Am J Med 1988; 85:
172-6
26 Gradon JD, Timpone JG, Schnittman SM. Emergence of unusual opportunistic pathogens in
AIDS: a review. Chn Infect Dis 1992; 15: 134-57
27 Harvey RL, Sunstrum JC. Rhodococcus equi infection in patients with and without HTV
infection. Rev Infect Dis 1991; 13: 139^5
28 Gilks CF, Brindle RJ, Otieno LS et al. Life-threatening bacteraemia in HTV-1 seropositive adults
admitted to hospital in Nairobi, Kenya. Lancet 1990; 336: 545-9
29 Gilks CF, Adam A, Batchelor B et al. A stable spectrum of bacteraemia and mycobacteraemia
in HIV-infected adults admitted to hospital in Nairobi. Submitted
30 Taelman H, Bogaerts J, Batungwanayo J et al. Community-acquired bacteraemia, fungaemia
and parasitaemia in febrila adults infected with HTV in Central Africa. Vth Int Conf on AIDS
in Africa, Kinshasa, October 1990 (FOD 1)
392
British Medical Bulletin 1998;54 (No. 2)
Acute bacterial infections and HIV disease
31 Vugia DJ, Kiehelbauch JA, Yeboue K et al. Pathogens and predictors of fatal septicaemia
associated with HIV infection in Ivory Coast, West Africa. / Infect Dis 1993; 168: 564-70
32 Grant AD, Djomand G, Smets P et al. Profound immunosuppression across the spectrum of
opportunistic disease among hospitalized HTV-infected adults in Abidjan, Core d'lvoire. AIDS
1997; 11: 1357-64
33 Gilks CF, Ojoo SA, Ojoo JC et al. Invasive pneumococcal disease in a cohort of predominantly
HIV-1 infected female sex workers in Nairobi, Kenya. Lancet 1996; 347: 718-23
34 French N, Mujugira A, Janoff EN, Gilks CF. Mortality and febrile illnesses among patients with
HTV-1 disease in Uganda and association with immunosuppression. ICAAC 1997; 37: 1-155
35 World Health Organization. AIDS. Interim proposal for a WHO staging system for HTV
infection and disease. Wkly Epidemiol Rec 1990; 65: 221-8
36 Tembo G, Friesan H, Asiimwe-Okiror G et al. Bed occupancy due to HTV/AIDS in an urban
hospital medical ward in Uganda. AIDS 1994; 8: 1169-71
37 Hassig SF, Perriens J, Baenda D et al. An analysis of the economic impact of HTV infection
among patients at Mama Yemo hospital, Kinshasa, Zaire. AIDS 1990; 4: 883-7
38 Gilks CF. What use is a clinical case definition for AIDS in Africa. BMJ 1991; 303: 1189-90
39 Gilks CF, Floyd K, Otieno LS et al. Some effects of a rising case-load of adult HTV-related
disease on a hospital in Nairobi. / Acquir Immune Def Syndr 1998; In press.
40 Morgan D, Maude GH, Malamba SS et al. HTV-1 disease progression and AIDS-defining
disorders in rural Uganda. Lancet 1997; 350: 245-50
41 Del Amo J, Petruckevitch A, Philips AN et al. Spectrum of disease in Africans with AIDS in
London. AIDS 1996; 10: 1563-9
42 Lucas SB, Hounnou A, Peaqcock C et al. The mortality and pathology of HIV infection in a
West African city. AIDS 1993; 7: 1569-79
43 Janoff EN, Obrien J, Thompson P et al. Streptococcus pneumomae colonisation, bacteraemia
and immune response among persons with HIV infection. / Infect Dis 1993; 167; 49-56
44 Cruciani M, Luzzati R, Fioredda F et al. Mucosal colonisation by pyogenic bacteria among
children with HIV infection. AIDS 1993; 7: 1533-34
45 Hirschtick RE, Glassroth J, Jordan MC et al. Baaerial pneumonia in persons with HTV. N EngI
] Med 1995; 333: 845-51.
46 Paul J, Kimari J, Gilks CF. Streptococcus pneumoniae resistant to penicillin and tetracycline
associated with HIV seropositivity. Lancet 1994; 346: 1034-5
47 Gotuzzo E, Frisancho O, Sanchez J et al. Association between AIDS and infection with
Salmonella typhi or Salmonella paratyphi in an endemic typhoid area. Arch Intern Med 1991;
151: 381-2.
48 Mivachari C, Batchelor B, Paul J, Wauyaki P, Gilks CF. Chronic diarrhoea among HTV-infected
adults in Nairobi, Kenya. / Infection 1998; In press
British Medical Bulletin 1998;54 (No. 2)
393