Complete thesisOpmaakV7

HIV and Mortality,
Autopsy Studies from Uganda
Dissertation submitted for the degree of Doctor
in Medical Sciences at the University of Antwerp
to be defended by
Janneke Cox
JannekeCox
Promotor
Robert Colebunders, University of Antwerp
HIVandMortality,AutopsyStudiesfromUganda
Institute of Tropical Medicine, Antwerp, Belgium
Infectious Diseases Institute, Kampala, Uganda
Antwerp 2014

2
HIV and Mortality, Autopsy studies from Uganda
Printed by: Provo in Gierle
Photography cover: Liz Cantu Davenport
The studies were financially supported by the Directorate General for Development
Cooperation (DGDC) through the Flemish Interuniversity council (VLIR-UOS), the
Nederlandse Vereniging van HIV-behandelaren, the KNCV Tuberculose fonds,
Boehringer Ingelheim B.V..
A travel grant was awarded by the Fonds Wetenschappelijk Onderzoek Flanders.
A travel and research grant was awarded by Gilead-Accordia Global Health
Foundation-IDSA-HIVMA.

3
HIV and Mortality,
Autopsy Studies from Uganda
Dissertation to be submitted for the degree of Doctor in Medical Sciences at
the University of Antwerp to be defended by
Proefschrift voorgelegd tot het behalen van de graad van Doctor in de
Medische Wetenschappen aan de Universiteit Antwerpen te verdedigen door
Janneke Cox
Infectious Diseases Institute, Kampala, Uganda
Promotor: Robert Colebunders, University of Antwerp
Antwerp, 14 november 2014

5
“To help the living, one must study the dead”.
Prabhat Jha, initiator of the Million Death study in India

7
Table of contents
Summary - Samenvatting............................................................................................................................9
Chapter 1 .........................................................................................................................................................15
Introduction
Chapter 2 .........................................................................................................................................................25
Autopsy Causes of Death in HIV-positive Individuals in sub-Saharan Africa and
Correlation with Clinical Diagnoses; an Overview of the Literature
Chapter 3 .........................................................................................................................................................47
Temporal Trends in All Cause Mortality in Adults attending an Urban HIV-clinic in
Uganda: a Retrospective Chart-review
Chapter 4 .........................................................................................................................................................67
Autopsy Acceptance Rate and Reasons for Decline in Mulago Hospital, Kampala,
Uganda
Chapter 5 .........................................................................................................................................................77
An Autopsy Study describing Causes of Death and comparing Clinico-Pathological
Findings among Hospitalized Patients in Kampala, Uganda
Chapter 6 .........................................................................................................................................................97
Needle Autopsy to Establish the Cause of Death in HIV-infected Hospitalized Adults in
Uganda: A Comparison to Complete Autopsy
Chapter 7 ...................................................................................................................................................... 117
Practice of Percutaneous Needle Autopsy; a Descriptive Study reporting Experiences
from Uganda
Chapter 8 ...................................................................................................................................................... 133
Discussion
Acknowledgements - Publication List - Curriculum Vitae...................................................... 147

11
Summary
Accurate knowledge of the causes of death is important to guide health care policy
and research, to improve the quality of care and to inform relatives and physicians.
Various methods exist to establish the cause of death, but each has limitations in
terms of feasibility, acceptability, accuracy and/or availability.
In sub-Saharan Africa (SSA) HIV is a significant cause of mortality, despite the fact that
it can be treated effectively with antiretroviral therapy (ART). Since 2003, ART has
become more widely available in SSA. This has decreased mortality in HIV-infected
patients, however in 2011 HIV still caused an estimated 1.2 million deaths.
To further decrease mortality in HIV-infected patients an adequate understanding of
the cause of death is vital. Therefore feasible, acceptable and accurate methods to
establish the cause of death are needed. In this thesis, we describe the causes of death
among HIV-infected patients in different settings in SSA and study the acceptability,
feasibility and accuracy of several methods to establish the death causes.
A review of complete and partial autopsies performed on HIV-infected adults in SSA
between 1993 and 2008 revealed tuberculosis (TB) to be the main pathologic finding
and cause of death. Infectious diseases caused the majority of deaths, with the lungs
and the central nervous system as main locations of disease. None of the patients
included in these studies had been treated with ART (Chapter 2).
In an urban cohort of 1028 deceased HIV-infected adults who attended an HIV-clinic
in Kampala, Uganda, we found that over the past 11 years, the proportion of deaths
from communicable conditions and AIDS-defining malignancies had decreased and
the proportion of deaths from chronic and other non-communicable conditions had
increased. The proportion of deaths on ART increased from 0% to 61% during that
same period. Nevertheless, in 2012 the majority of deaths were still due to
communicable conditions and AIDS-defining malignancies (Chapter 3).
Two autopsy series performed among hospitalized HIV-infected adults at Mulago
Hospital in Kampala in 2009 and 2013 revealed infectious diseases as the cause of
death in 83%-91% of patients. TB was the main infectious cause in both series. ART
use in both studies was 27% and 61% respectively and the median CD4 count 50 and
47 cells/mm3 (Chapter 5 and 6).
We concluded that infections continue to cause the majority of deaths of both in- and
outpatient HIV-infected adults in SSA, despite ART expansion.
When assessing the accuracy of the pre-mortem clinical diagnosis in HIV-infected
hospitalized adults, we found a low concordance compared to complete autopsy-
established diagnosis. Based on the hospital chart, only 12% of the autopsy-

12
established diagnoses were confirmed while the patient was alive. Another 27% of
diagnoses were highly suspected and 16% were considered. Main reasons for not
considering a diagnosis pre-mortem were dual infections or atypical disease
presentation. Although numbers are small, this suggests that mortality data based on
clinical diagnosis are not very reliable. Moreover, diseases not considered clinically
will often not be treated (Chapter 5).
Complete autopsy is considered the gold standard to establish cause of death.
Nevertheless, complete autopsies are infrequently performed for a variety of reasons,
including low acceptance rates by relatives. We found an autopsy acceptance rate of
38% in our 2009 autopsy series. We assessed the reason for decline among non-
accepters and found that the most commonly stated reason was not wanting to delay
the funeral. With timely requests and well-organized, fast autopsy procedures,
autopsy acceptance rates may therefore be increased (Chapter 4).
Even in cases where complete autopsies are accepted by the family, they are not
always feasible. Moreover, they pose a risk for health care workers, especially if health
care facilities are poorly equipped. To address this challenge, we evaluated needle
autopsies as an alternative method. To compensate for the loss of visual judgment
during the procedure, we added ultrasound guidance. We subsequently performed a
blind needle autopsy, an ultrasound guided needle autopsy and a complete autopsy in
the same patient and compared the yield of the first 2 procedures against complete
autopsy. The concordance for all major diagnoses between the blind needle autopsy
and the complete autopsy was 50%. For TB, the main major diagnosis, concordance
was significantly higher (73%). We found that the addition of ultrasound guidance did
not significantly increase concordance (Chapter 6). Despite a lack of experience with
needle autopsies among our study team, we obtained adequate tissue for further
histological evaluation for the different organs in 56-99% of cases. Moreover, a clear
learning curve was observed (Chapter 7). This suggests that needle autopsy could be a
reasonably accurate and feasible alternative to complete autopsy.
Since the introduction of ART, mortality in HIV-infected patients in SSA has decreased.
However, the causes of death have hardly changed in the populations we studied:
treatable HIV-related infections and malignancies continue to cause the majority of
deaths. Adequate knowledge on the causes of death can inform and evaluate
interventions to decrease mortality. Therefore causes of death and the methods to
establish these need to be studied.

13
Samenvatting
Betrouwbare kennis over doodsoorzaken is van belang om richting te geven aan
gezondheidsbeleid, om kwaliteit van zorg te evalueren en om familieleden en artsen te
informeren. Er zijn verschillende methoden beschikbaar om de doodsoorzaak vast te
stellen. Iedere methode heeft echter beperkingen wat betreft uitvoerbaarheid,
acceptatie, betrouwbaarheid en beschikbaarheid.
HIV is een belangrijke doodsoorzaak in sub-Sahara Afrika (SSA). Met antiretrovirale
therapie (ART) kan HIV behandeld worden. Sinds 2003 is de beschikbaarheid van ART
in SSA enorm toegenomen. Dat heeft er onder andere toe geleid dat de mortaliteit van
HIV-geïnfecteerde patiënten gedaald is. Echter, in 2011 veroorzaakte HIV nog steeds
1.2 miljoen doden.
Om mortaliteit in HIV-geïnfecteerde patiënten te verminderen is betrouwbare kennis
van de onderliggende doodsoorzaken nodig. Daarom is het belangrijk om goede
methoden te hebben om de doodsoorzaak vast te stellen. In dit proefschrift
beschrijven we de doodsoorzaken van HIV-geïnfecteerde volwassenen in SSA en
bestuderen de acceptatie, betrouwbaarheid en uitvoerbaarheid van verschillende
methoden die gebruikt kunnen worden om de doodsoorzaak vast te stellen.
In een overzichtsartikel van volledige- en partiële autopsie studies in HIV-
geïnfecteerden in SSA, uitgevoerd tussen 1993 en 2008, bleek dat tuberculose (TB) de
belangrijkste ziekte en doodsoorzaak bij volwassenen is. De meerderheid van de
sterfgevallen werd veroorzaakt door infecties, waarbij de longen en het centraal
zenuwstelsel het vaakst betrokken waren. De besproken studies bevatten geen
patiënten die behandeld werden met ART (Hoofdstuk 2).
We bestudeerden retrospectief 1028 HIV-geïnfecteerde volwassenen die overleden
tussen 2002 en 2012 terwijl ze in zorg waren in een polikliniek in Kampala. De
proportie doden ten gevolge van overdraagbare ziekten en AIDS-gerelateerde
maligniteiten nam af in de tijd en de proportie doden ten gevolge van chronische en
andere niet overdraagbare aandoeningen nam toe. In dezelfde tijdspanne steeg de
proportie doden op ART van 0% naar 61%. Echter, ook in 2012 stierf de meerderheid
aan een overdraagbare aandoening of AIDS-gerelateerde maligniteit (Hoofdstuk 3).
In de 2 autopsie studies in Mulago ziekenhuis in Kampala uit 2009 en 2013 overleed
83-91% van de HIV-geïnfecteerde volwassen aan een infectie. TB was de belangrijkste
doodsoorzaak in beide studies. Het aantal patiënten op ART was respectievelijk 27%
en 61% en het mediane CD4 aantal 50 en 47 cellen/mm3(Hoofdstuk 5 en 6).
Kortom, de meerderheid van de HIV-geïnfecteerde volwassenen in SSA overlijdt nog
steeds aan infecties, zowel in de klinische als poliklinische setting en ondanks de
toegenomen beschikbaarheid van ART.

14
We vergeleken de klinische diagnose met de autopsie diagnose en vonden een lage
overeenkomst tussen beiden; slechts 12% van de autopsie diagnoses was klinisch
bevestigd. In 27% was er een hoge klinische verdenking en in 16% werd de diagnose
overwogen. De belangrijkste redenen om een diagnose bij leven niet te overwegen
waren een atypische ziekte presentatie of multipele diagnoses in 1 patiënt. Ondanks
de lage aantallen in deze studie lijkt het erop dat mortaliteitsdata die gebaseerd is op
de klinische diagnose niet erg betrouwbaar is. Daarnaast is het onwaarschijnlijk dat
een ziekte die klinisch niet overwogen wordt, behandeld zal worden (Hoofdstuk 5).
Volledige autopsie is de gouden standaard voor het vaststellen van de doodsoorzaak.
Echter volledige autopsies worden niet vaak uitgevoerd. Een van de redenen hiervoor
is het weigeren van een autopsie door naasten. In onze studie uit 2009 werd in 38%
van de gevallen een volledige autopsie geaccepteerd. Voor degenen die de volledige
autopsie weigerden bleek het niet willen vertragen van de uitvaart de belangrijkste
reden. Indien tijdig om toestemming voor de autopsie verzocht wordt en vervolgens
de autopsie vlot plaatsvindt, kan de acceptatiegraad van autopsies wellicht verhoogd
worden (Hoofdstuk 4).
Ook wanneer volledige autopsies geaccepteerd worden, zijn ze niet altijd uitvoerbaar.
Daarnaast is een autopsie risicovol voor degene die de procedure uitvoert, vooral in
omstandigheden waar gezondheidscentra matig toegerust zijn. Dat was voor ons de
reden om een alternatieve methode, naaldautopsie met of zonder echogeleide, nader
te bestuderen. De overeenstemming voor majeure diagnoses tussen de blinde
naaldautopsie en de volledige autopsie was 50%. Voor TB, de meest voorkomende
diagnose was de overeenkomst significant hoger (73%). De toevoeging van echo
leidde niet tot significant betere overeenstemming (Hoofdstuk 6). Ondanks een
gebrek aan ervaring bij degenen die de naaldautopsies uitvoerden werd er van de
verschillende organen in 56-99% weefsel verkregen van voldoende kwaliteit voor
nader histologisch onderzoek. Daarnaast was er sprake van een duidelijke leercurve
bij de uitvoerders (Hoofdstuk 7). Zodoende lijkt naaldautopsie een redelijk
betrouwbaar en haalbaar alternatief voor volledige autopsie.
De mortaliteit van HIV-geïnfecteerde patiënten in SSA is gedaald sinds de introductie
van ART. Echter, de doodsoorzaken zijn grotendeels dezelfde gebleven, ten minste in
de populaties die wij bestudeerden: de meerderheid van de patiënten overlijdt ook nu
nog aan behandelbare, HIV-gerelateerde aandoeningen. Betrouwbare kennis van de
doodsoorzaken is van belang om interventies om mortaliteit te verlagen te
identificeren en te evalueren. Daarom is het belangrijk doodsoorzaken en de
methoden om deze vast te stellen te bestuderen.

17
Introduction
Why wanting to know the cause of death?
Having a good understanding of the causes of death within a population is of great
importance. On a macro-level this information is needed to design effective healthcare
programs, assign health funding and evaluate the impact of interventions in the
healthcare system1. Moreover, it is needed to define health-related research priorities
and detect newly emerging disease trends and epidemics. On an individual level,
reliable knowledge on a cause of death is used to evaluate quality of care, improve the
understanding of diseases, offer consolation to relatives and provide information
regarding heritable or contagious diseases2.
Establishing the cause of death
Several methods are used to establish the cause of death (Box 1). The oldest method is
autopsy, opening the body to examine the organs and collect tissue and/or bodily
fluids3. Although still the undisputed gold standard, its practice is limited due to the
need for well-trained staff and well-equipped health facilities, its cost, its mutilating
nature, declining acceptance rates and the risk of disease transmission among others2.
Less invasive methods are known
as partial autopsy or minimal
invasive autopsy. This includes
opening only a part of the body to
examine and obtain tissue of (a)
certain organ(s) and the use of
percutaneous needles to obtain
tissue. Studies that have reported
upon these techniques focused
mainly on certain populations e.g.
children or certain diseases e.g.
pulmonary diseases4-7.
More of recent date, post-mortem
imaging is reported as an
alternative method to establish
the cause of death. The main
advantages are the safety of the
procedure and the non-invasive
nature. Both in adults and
children, different techniques (including MRI and CT) have been studied and
reasonable test characteristics are reported8,9. An important disadvantage of these
imaging techniques however is their limited availability in many parts of the world.
Box 1. Methods to establish the cause of
death.
Complete autopsy: sectioning the body to examine
all organs including the brain and collect tissue for
further histological assessment. Synonym: necropsy.
Partial autopsy: sectioning (a) certain part(s) of the
body to examine (a) certain organ(s) and collect
tissue of the(se) organ(s) for further histological
assessment.
Minimal invasive autopsy: obtain tissue(s) of (a)
certain organ(s) without sectioning the body e.g. by
using a percutaneous biopsy needle.
Postmortem imaging: scanning the body
postmortem using CT or MRI technology.
Death certification: statement by the treating
physician of the presumed cause of death and any
underlying, contributing diseases.
Verbal autopsy: interviewing the primary care taker
of the deceased to gather information on the
circumstances at the moment of death. The cause of
death is assigned based on review of the information
by health care workers or an automated computer
system.

18
Also, their accuracy for certain highly prevalent diseases like malaria or infectious
diarrhea still needs to be determined.
Death certificates completed by health care workers is another method of assigning
the cause of death. The accuracy of this methods greatly depends on the diagnostic
modalities available pre-mortem, skills of individual physicians and legal and
institutional frameworks10. Moreover, in 2005 only 36% of all deaths worldwide
occurred within healthcare facilities and where certified, with large variation between
different regions11.
To ascertain more information on uncertified deaths, verbal autopsy has been
developed. Trained field workers interview the primary care taker of the deceased
within 6 months after the death occurred using a predefined questionnaire. Several
validated verbal autopsy tools are available, however, their use is not undisputed12.
Certain diseases and disease-categories miss distinctive traits, which makes them
difficult to diagnose with a verbal autopsy. Malaria in adults is a well-known and
heavily debated example1. Moreover, until now, verbal autopsies have only been
validated using death certificates as gold standard. The accuracy of death certificates
is highly variable and its use as gold standard therefore questionable10.
Causes of death globally
The global burden of disease study estimates global disease specific causes of death.
For each country, they collect and combine the available information acquired with
the different methods described above. Statistical models are used to fill in the gaps.
In 2010 they estimated a total of 52.8 million deaths worldwide. The 3 major causes of
death were ischemic heart diseases, stroke and chronic obstructive pulmonary
disease. Ischemic heart disease, lower respiratory tract infections and
cerebrovascular disease accounted for the most years of life lost globally13.
Causes of death in sub-Saharan Africa, the role of HIV
In 2002, sub-Saharan Africa (SSA) had an estimated total population of almost 700
million, approximately 11% of the world population14. In that same year the total
estimated number of deaths in SSA was 10.8 million, slightly less than 20% of the total
global mortality at that time15. Forty-six percent of deaths occurred in the age group
below 15 years and 36% in the age group between 15 and 59 years. The main causes
of death were HIV, malaria and lower respiratory infections accounting for resp. 20%,
10% and 10% of deaths15. Since 2005, the number of HIV-related deaths is declining
in SSA as a result of improved treatment and care for HIV-infected patients16-19.
Nevertheless, HIV/AIDS was one of the leading causes of years of life lost in SSA
according to the 2010 global burden of disease study13. The World Health
Organization (WHO) estimated that 1.2 million HIV-related deaths occurred in SSA in
2011, almost 71% of the global total20.

19
HIV, opportunistic infections and infection related malignancies
HIV-infection itself leads seldom to death. In the majority of cases it is the
deteriorating effect of HIV on the immune system that renders the patient susceptible
for opportunistic infections and certain malignancies, which eventually cause HIV-
infected patients to die21. Well-known examples are Pneumocystis jirovecii infection
causing pulmonary infection and Kaposi's sarcoma, a malignancy caused by a human
herpes virus 8. The time course between the primary HIV infection and the decline of
the immune system to a level that opportunistic infections can occur varies
considerably from patient to patient, but is in most cases 7-10 years22.
Antiretroviral therapy
Antiretroviral drugs oppose the immune-modulating effects of HIV by inhibiting the
viral replication22,23. In the late nineties, it became clear HIV is best treated with a
combination of at least 3 of these antiretroviral drugs, known as antiretroviral
therapy (ART)24,25. By starting ART timely the occurrence of opportunistic infections
can be prevented and death averted. With adequate and timely ART, HIV-infected
patients can have a life expectancy that is not very different from the HIV-uninfected
population26,27.
Causes of Death in HIV-infected patients in sub-Saharan Africa
When describing causes of mortality in HIV-infected patients, a distinction needs to be
made between mortality in patients not on ART, early mortality in the first period
after the start of ART and mortality in patients that are (successfully) treated with
ART28.
The main causes of death in untreated HIV-infected patients are infections with
Mycobacterium tuberculosis, yeasts and bacteria that lead to meningitis, pneumonia or
sepsis29. When patients are started on ART, high mortality rates in the first 12 months,
with a peak in the first 3 months are observed28,30. The main causes of these early
deaths include tuberculosis (TB), advanced HIV-infection, chronic diarrhea and
cryptococcal meningitis30. So, the infections causing early mortality are more or less
the same infections that cause death in untreated HIV-infected patients. This reflects
the poor immune status of patients starting ART. Treatment related-events like the
immune reconstitution syndrome (IRIS) after the start of ART cause very limited
mortality and mainly in specific disease conditions31,32. Also, ART related toxicity does
not seem an important cause of mortality33.
Once patients have survived the first months on ART and restoration of the immune-
system occurs, mortality drops substantially, also in Africa34,35. Epidemiological data
from African surveillance sites 4-5 years after the introduction of ART show this drop
sustains and is due to a decrease in HIV-related communicable diseases17,18.

20
Information on cause-specific mortality in long-term treated HIV-infected patients is
scarce. Case series, describing a maximum of 37 patients, report virologic and
immunologic failure and death of HIV-related causes in 70-100% of patients on ART
for more than 1 year36,37. However, it is not clear if and how these data represent a
larger population.
HIV in Uganda
Uganda was among the first countries to describe 'slim disease', a condition that was
later to be found HIV38. The HIV prevalence estimate in Uganda was 11% in 1992 but
has decreased to 6.7% in 2011. Nevertheless, the absolute number of HIV-infected
individuals is still increasing due to population growth in the last 2 decades39.
Free ART became available in Ugandan public health-facilities in 2003, thanks to
international funding through the Multi-Country HIV/AIDS Program (2002), the
Global Fund to fight AIDS, Tuberculosis and Malaria (2003) and the President's
Emergency Plan for AIDS Relief (2004)40. Since then, ART use has been scaled up
rapidly. In 2012, ART coverage was 67% (95% confidence interval 57-71), using the
350 cells/mm3 threshold19.
For patients on ART, early mortality rates of 11.8-17.9/100 person years (PY) at risk
are reported. A steep decrease in mortality rates is observed when ART is continued;
2.3-2.4/100 PY at risk in the second and 1.2/100PY at risk in the third year after start
of ART41,42. The available reports on causes of death within HIV-infected adults are
based on medical certificates or verbal autopsy data. The main causes of death for
early mortality are TB (11-57%), central nervous system infections (13-25%) and
chronic diarrhea (5-13%)41-43. Mortality data for patients on ART beyond one year are
scarce.
Pathology services in Uganda
The access to pathology services in many African countries is limited44. Uganda had in
2010 an estimated population of more than 33 million45. Currently only 18 trained
pathologists are actively working of which all but one work within Kampala (personal
communication R. Lukande). In comparison, in Northern America and Europe, in
general 14-40 pathologists are working per million population44.

21
Objectives and Outline
As outlined above, it is of great importance to have accurate information on the causes
of death. The different methods to establish the cause of death each have their
shortcomings regarding feasibility, accessibility and accuracy among others. Mortality
and causes of death in HIV-infected patients depend on the use of adequate ART.
Increased availability of ART in SSA has decreased mortality in HIV-infected patients.
Nevertheless, HIV is still a major contributor to death in SSA.
Against this background we have performed our studies which had a twofold overall
objective; to describe the causes of death and to study methods to establish the cause
of death in HIV-infected patients in SSA.
More specifically, the objectives of this thesis are as follows:
Causes of death
- To review and describe the autopsy-established causes of death in HIV-infected
patients in sub-Saharan Africa (Chapter 2).
- To evaluate and describe the causes of death in a large urban HIV-treatment clinic
over time (Chapter 3).
- Describe the autopsy confirmed causes of death in HIV-infected and uninfected
adults that die during hospitalization at a tertiary hospital (Chapter 5 and 6).
Feasibility and accuracy of different methods to establish the cause of
death
- To evaluate the acceptability and the reason for decline of complete autopsies
(Chapter 4).
- To compare clinical to autopsy-established causes of death in hospitalized HIV-
infected and uninfected adults (Chapter 5).
- To assess the accuracy of needle autopsies to establish a cause of death in
hospitalized HIV-infected adults (Chapter 6).
- To describe the practicalities of needle autopsies (Chapter 7).
All studies were conducted in Kampala, the capital of Uganda. The institutions where
we carried out our studies, Mulago hospital and the Reach Out Mbuya Parish
HIV/AIDS initiative, are described in more detail in the next chapters.

22
References
1. Vogel G. Global health. How do you count the dead? Science. Jun 15
2012;336(6087):1372-1374.
2. Burton JL, Underwood J. Clinical, educational, and epidemiological value of autopsy.
Lancet. Apr 28 2007;369(9571):1471-1480.
3. Clark MJ. Autopsy. Lancet. Nov 19 2005;366(9499):1767.
4. Wellmann K. The needle autopsy. A retrospective evaluation of 394 consecutive cases.
Am J Clin Pathol. Oct 1969;52(4):441-444.
5. Murray J, Sonnenberg P, Nelson G, Bester A, Shearer S, Glynn JR. Cause of death and
presence of respiratory disease at autopsy in an HIV-1 seroconversion cohort of southern
African gold miners. Aids. Nov 2007;21 Suppl 6:S97-S104.
6. Abouya YL, Beaumel A, Lucas S, et al. Pneumocystis carinii pneumonia. An uncommon
cause of death in African patients with acquired immunodeficiency syndrome. Am Rev Respir
Dis. Mar 1992;145(3):617-620.
7. Garg S, Punia RP, Basu S, Mohan H, Bal A. Comparison of needle autopsy with
conventional autopsy in neonates. Fetal Pediatr Pathol. 2009;28(3):139-150.
8. Thayyil S, Sebire NJ, Chitty LS, et al. Post-mortem MRI versus conventional autopsy in
fetuses and children: a prospective validation study. Lancet. Jul 20 2013;382(9888):223-233.
9. Roberts IS, Benamore RE, Benbow EW, et al. Post-mortem imaging as an alternative to
autopsy in the diagnosis of adult deaths: a validation study. Lancet. Jan 14 2012;379(9811):136-
142.
10. Shojania KG, Burton EC, McDonald KM, Goldman L. Changes in rates of autopsy-
detected diagnostic errors over time: a systematic review. JAMA. Jun 4 2003;289(21):2849-
2856.
11. Mathers CD, Fat DM, Inoue M, Rao C, Lopez AD. Counting the dead and what they died
from: an assessment of the global status of cause of death data. Bull World Health Organ. Mar
2005;83(3):171-177.
12. Bassat Q, Ordi J, Vila J, et al. Development of a post-mortem procedure to reduce the
uncertainty regarding causes of death in developing countries. The Lancet Global Health.
2013;1(3):e125-126.
13. Lozano R, Naghavi M, Foreman K, et al. Global and regional mortality from 235 causes
of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of
Disease Study 2010. Lancet. Dec 15 2012;380(9859):2095-2128.
14. International Database. United States Department of Commerce.
http://www.census.gov/population/international/data/idb/informationGateway.php.
15. Rao C LA, Hemed Y. Causes of Death. Disease and Mortality in Sub-Saharan Africa. 2nd
edition ed. Washington (DC): World Bank; 2006.
16. Brinkhof MW, Boulle A, Weigel R, et al. Mortality of HIV-infected patients starting
antiretroviral therapy in sub-Saharan Africa: comparison with HIV-unrelated mortality. PLoS
Med. Apr 28 2009;6(4):e1000066.
17. Chihana M, Floyd S, Molesworth A, et al. Adult mortality and probable cause of death in
rural northern Malawi in the era of HIV treatment. Trop Med Int Health. Aug 2012;17(8):e74-83.
18. Herbst AJ, Mafojane T, Newell ML. Verbal autopsy-based cause-specific mortality
trends in rural KwaZulu-Natal, South Africa, 2000-2009. Popul Health Metr. 2011;9:47.
19. World Health Organisation. Global update on HIV treatment 2013: results, impact and
opportunities. Geneva 2013.
20. World Health Organisation. Estimated adult and child death from AIDS 2011. Geneva
2012.
21. Dan Longo AF, Dennis Kasper, Stephen Hauser, J. Jameson, Joseph Loscalzo. Harrison's
Principles of Internal Medicine. 2011.
22. R. Dolin HM, M. Saag AIDS Therapy. Churchill Livingstone, Elsevier Inc. ; 2008.
23. Hirschel B, Francioli P. Progress and problems in the fight against AIDS. N Engl J Med.
Mar 26 1998;338(13):906-908.

23
24. Gulick RM, Mellors JW, Havlir D, et al. Treatment with indinavir, zidovudine, and
lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral
therapy. N Engl J Med. Sep 11 1997;337(11):734-739.
25. Hammer SM, Squires KE, Hughes MD, et al. A controlled trial of two nucleoside
analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell
counts of 200 per cubic millimeter or less. AIDS Clinical Trials Group 320 Study Team. N Engl J
Med. Sep 11 1997;337(11):725-733.
26. Samji H, Cescon A, Hogg RS, et al. Closing the Gap: Increases in Life Expectancy among
Treated HIV-Positive Individuals in the United States and Canada. PLoS One.
2013;8(12):e81355.
27. Mills EJ, Bakanda C, Birungi J, et al. Life expectancy of persons receiving combination
antiretroviral therapy in low-income countries: a cohort analysis from Uganda. Ann Intern Med.
Aug 16 2011;155(4):209-216.
28. Lawn SD, Harries AD, Anglaret X, Myer L, Wood R. Early mortality among adults
accessing antiretroviral treatment programmes in sub-Saharan Africa. Aids. Oct 1
2008;22(15):1897-1908.
29. Lucas SB, Hounnou A, Peacock C, et al. The mortality and pathology of HIV infection in
a west African city. Aids. Dec 1993;7(12):1569-1579.
30. Gupta A, Nadkarni G, Yang WT, et al. Early mortality in adults initiating antiretroviral
therapy (ART) in low- and middle-income countries (LMIC): a systematic review and meta-
analysis. PLoS One. 2011;6(12):e28691.
31. Mateen FJ, Nath A. Central nervous system-immune reconstitution inflammatory
syndrome in resource-limited settings: current burden and future needs. Aids. Sep 24
2012;26(15):1851-1855.
32. Worodria W, Menten J, Massinga-Loembe M, et al. Clinical spectrum, risk factors and
outcome of immune reconstitution inflammatory syndrome in patients with tuberculosis-HIV
coinfection. Antivir Ther. 2012;17(5):841-848.
33. Kovari H, Sabin CA, Ledergerber B, et al. Antiretroviral drug-related liver mortality
among HIV-positive persons in the absence of hepatitis B or C virus coinfection: the data
collection on adverse events of anti-HIV drugs study. Clin Infect Dis. Mar 2013;56(6):870-879.
34. De Luca A, Marazzi MC, Mancinelli S, et al. Prognostic value of virological and
immunological responses after 6 months of antiretroviral treatment in adults with HIV-1
infection in sub-Saharan Africa. J Acquir Immune Defic Syndr. Mar 1 2012;59(3):236-244.
35. Kasamba I, Baisley K, Mayanja BN, Maher D, Grosskurth H. The impact of antiretroviral
treatment on mortality trends of HIV-positive adults in rural Uganda: a longitudinal population-
based study, 1999-2009. Trop Med Int Health. Aug 2012;17(8):e66-73.
36. Wong EB, Omar T, Setlhako GJ, et al. Causes of death on antiretroviral therapy: a post-
mortem study from South Africa. PLoS One. 2012;7(10):e47542.
37. Karstaedt AS. Profile of cause of death assigned to adults on antiretroviral therapy in
Soweto. South African medical journal = Suid-Afrikaanse tydskrif vir geneeskunde. Aug
2012;102(8):680-682.
38. Serwadda D, Mugerwa RD, Sewankambo NK, et al. Slim disease: a new disease in
Uganda and its association with HTLV-III infection. Lancet. Oct 19 1985;2(8460):849-852.
39. Uganda AIDS Commission. Global AIDS Response Progress Report: Uganda. Kampala:
Uganda AIDS Commission;2012.
40. Ministry of Health U.ganda Antiretroviral Therapy. Treatment guidelines Uganda.
2004.
41. Castelnuovo B, Manabe YC, Kiragga A, Kamya M, Easterbrook P, Kambugu A. Cause-
specific mortality and the contribution of immune reconstitution inflammatory syndrome in the
first 3 years after antiretroviral therapy initiation in an urban African cohort. Clin Infect Dis. Sep
15 2009;49(6):965-972.
42. Amuron B, Levin J, Birunghi J, et al. Mortality in an antiretroviral therapy programme
in Jinja, south-east Uganda: a prospective cohort study. AIDS Res Ther. 2011;8:39.
43. Kyeyune R, den Boon S, Cattamanchi A, et al. Causes of early mortality in HIV-infected
TB suspects in an East African referral hospital. J Acquir Immune Defic Syndr. Dec
2010;55(4):446-450.

24
44. Benediktsson H, Whitelaw J, Roy I. Pathology services in developing countries: a
challenge. Arch Pathol Lab Med. Nov 2007;131(11):1636-1639.
45. United Nations World Population Prospects. 2010.

25
Chapter 2
Autopsy Causes of Death in HIV-positive
Individuals in sub-Saharan Africa and Correlation
with Clinical Diagnoses; an Overview of the
Literature
Janneke A. Cox1
, Robert L. Lukande2
, Sebastian Lucas3
, Ann M. Nelson4
, Eric
Van Marck5
, Robert Colebunders1,6
1
2
Department of Pathology, Makerere University, Kampala, Uganda
3
Department of Histopathology, King’s College London School of Medicine, London,
United Kingdom
4
Armed Forces Institute of Pathology, Washington, United States of America
5
Department of Pathology, University of Antwerp, Belgium
6
Epidemiology and Social Medicine, University of Antwerp, Belgium
AIDS Reviews 2010; 12(4): 183‐94.

27
Abstract
Despite the persistent high HIV-related mortality in sub-Saharan Africa, limited
information on causes of death is available. Pathological autopsies are the gold
standard to establish causes of death. In this review we describe the autopsy series
performed among HIV infected individuals in sub-Saharan Africa over the last 2
decades. We identified 9 complete and 11 partial or minimal invasive autopsy series.
Complete autopsies were performed in 593 HIV positive adults and 177 HIV positive
children. Post-mortem diagnoses were mainly infectious diseases. Tuberculosis was
the most frequent, present in 21-54% of HIV positive adults and was considered the
cause of death in 32-45%. Overall, pulmonary infections accounted for approximately
66% of pathology and central nervous system infections for approximately 20%. A
high discordance between clinical and post-mortem diagnoses was observed. This
review emphasises the need for reliable information on causes of death in order to
improve HIV patient care, guide further research and inform health policy.

28
Introduction
Since the worldwide roll out of anti-retroviral treatment (ART) mortality due to HIV
has dramatically decreased1,2. However substantial AIDS mortality persists; In 2008,
the number of deaths due to AIDS in sub-Saharan Africa was estimated 1.4 million3. A
large proportion of those are children and adults in sub-Saharan Africa who still lack
access to HIV treatment3. But also among those on ART, mortality remains high
despite good immunologic and virologic responses4. In this group most death occurs
in the first months after the start of ART5,6. This early mortality seems to be caused
mainly by HIV-related opportunistic infections and malignancies. The immune
reconstitution inflammatory syndrome (IRIS) and ART toxicity seem to be less
important causes of early death, but how much these conditions contribute to
mortality remains to be determined4,6,7.
Most information on the causes of HIV-related death in Africa comes from clinical
studies and studies using verbal autopsies4,8,9. However pathological autopsy, a
macro- and microscopic post mortem examination of the body, is the gold standard to
determine causes of death. Despite improved diagnostic facilities there is a high
discrepancy between clinical and post-mortem diagnosis. A study reviewing autopsies
done in HIV-positive individuals who died in the United Kingdom showed that in 70%
of autopsies the cause of death was different from the clinical diagnosis and that 36%
of the opportunistic infections were clinically missed10. In resource-limited settings
this discrepancy is expected to be even higher because of the limited access to pre-
mortem examinations.
A better knowledge of the causes of death is essential to improve clinical practice, to
guide further research and to inform policy makers. In this paper we review the
results of autopsy studies done in sub-Saharan Africa in HIV positive individuals and
the correlation between autopsy and clinical diagnoses.
Search strategy and definitions
The Medline electronic database was searched via PubMed without language or time
limitations. The following Medical Subject Headings (MESH) were used: “Africa South
of the Sahara”, “HIV”, “Acquired Immunodeficiency Syndrome” and “Autopsy”. This
returned 264 articles. Original studies describing complete autopsies, partial
autopsies and minimally invasive autopsies in at least 8 HIV positive individuals in
Sub Saharan Africa, were considered. Sub-studies of larger post-mortem series and
studies describing verbal autopsy results or clinical causes of death were not
considered for review. Reference lists of relevant papers were used to identify
additional publications. The abstracts (2001-2010) of the Conference on Retroviruses
and Opportunistic Infections (CROI) and of the International AIDS society conferences
were searched. Experts in the field were contacted to suggest any additional studies.

29
A complete autopsy was defined as a macro- and microscopical examination of all
thoracic and abdominal organs and the brain. A partial autopsy was defined as a
macro- and microscopic examination of a number of organs or a particular organ.
Minimally invasive autopsies included blind or targeted percutaneous needle biopsies
and examinations using endoscopic or laparoscopic techniques.
Literature overview
Description of the studies
We identified 19 reports and 1 abstract from 10 different African countries describing
9 complete and 11 partial or minimal invasive autopsy series. Table 1 summarises
study- and patient characteristics11-31. Ten of these autopsy series also included HIV
negative individuals (table 1)12-15,17,19,20,22,23,26. Seven of these studies compared
autopsy findings to clinical diagnosis (Table 2)11,14,21,23-26.
Complete autopsies were performed in 593 HIV positive adults and 177 HIV positive
children. Partial autopsies varied from sampling of lung tissue only to sampling all
organs except for the brain. Partial autopsies were chosen for various reasons. In the
Democratic Republic of Congo they were done because of poor acceptance of complete
autopsies24. Among South African miners they were done to evaluate lung pathology
for compensation purposes23. Studies using minimal invasive autopsies were only
conducted in children and designed to answer organ- or disease specific research
questions15,17,27,28,30.
Nearly all studies were performed in major tertiary-level health institutions and
among hospitalised patients. Some studies also included a limited number of dead
persons brought to the hospital from outside16,19,20. Kibayashi et al. selected HIV
suspects from a large sample of forensic autopsies18. Chakraborty et al. included non-
hospitalised orphans29. The focus of the studies varied widely, including patients from
very different hospital wards12,13,19,22,31. Two studies specifically stated they
prioritised patients with an unknown cause of death13,24.
In four studies no informed consent was required for the autopsy. In these studies
autopsy rates of 24-100% were reported12,19,20,29. Why these 4 studies did not have
autopsy rates of 100% was mainly because of capacity restraints, but a positive
outcome is that the cases examined could be more representative, ie unselected.
Menedez et al. consecutively included all maternal death that occoured in the hospital,
resulting in 78% coverage22. For the other studies the precise mode for inclusion was
not stated. The 3 studies from Ivory Coast were the only ones that included HIV
serotype 2 infected patients; 113 adults (45 HIV-2 and 68 dually reactive) and 2
children (both HIV-2 infected)12,19,20.
Inclusion periods mostly did not exceed 2 years and the number of autopsies varied
between 10 and 247. The median CD4 count in adults varied between 48-101

30
cells/L19,21,26. For children a mean CD4 percentage of 14,8% was reported28. The
median number of hospitalisation days before death was 4-6 days13,14,19,21 with one
study reporting a modal survival of 1 day (14% of deaths) (18). None of the studies
reported that patients received ART. Only Chakraborty et al. mentioned that all
children included in their post-mortem study had been on prophylactic co-
trimoxazole29.
All studies used haematoxylin and eosin (H&E) stains on all tissues and added Ziehl
Neelsen (ZN), Gram, Periodic acid Schiff and Grocott silver staining as appropriate.
Most studies also had immunohistochemical analysis at their disposal, with antibody
testing against cytomegalovirus (CMV), Toxoplasma gondii, Epstein Barr virus (EBV),
respiratory syncytial virus and adenovirus12-15,19-22,26,28,29. One study used
immunohistochemistry for lymphoma diagnostics19. For suspected tuberculosis (TB)
diagnostics some studies used beside ZN staining, cultures19,21,27 or polymerase chain
reaction21.
Pathology findings Adults
Infectious diseases
The overall majority of post-mortem diagnoses were infectious diseases (table 1). TB
was the most frequent diagnosis, observed in 21-54% of HIV positive adults and
considered as the cause of death in 32-45%11,13,19,23,24,26. TB was as common in
patients selected from pulmonology wards as in unselected hospitalised patients and
was disseminated in at least 80% of cases12,13,19,26. Martinson et al. autopsied 47 HIV
positive patients suspected of TB during hospitalisation. TB was confirmed in 79% of
these. Of these, 92% had disseminated TB. The remaining 21% had
bronchopneumonia (60%), CMV pneumonitis (50%) and Kaposi’s sarcoma (KS)
(20%)21.
Other important pulmonary infections were bacterial pneumonia (present in 23-
34%), interstitial pneumonia (present in 17%) and Pneumocystis jiroveci pneumonia
(PJP, present in 9-14%)11-13,19,23,24,26. Only one study reported nocardiosis, in 4% of
cases19,32. Overall, pulmonary infections were present in at least 2 out of 3 patients
and were the cause of death in nearly half of the patients.
The central nervous system was the second most affected organ-system by infections.
Bacterial meningitis, tuberculous meningitis, lymphocytic meningitis, cryptococcal
meningitis and cerebral toxoplasmosis were the main diagnoses, together accounting
for approximately 20% of the pathology13,19,26. Later re-examination of some cases
from the Ivory Coast adult series (17) with immunohistochemistry demonstrated both
pneumococcal and meningococcal meningitis infections (Lucas S,, personal
communication).

31
Non infectious diseases
Malignancies caused 11-16% of pathology. This consisted mainly of Kaposi’s sarcoma
(KS). Non-Hodgkin lymphoma (NHL) appeared rarely if at all in the studies13,24. A sub-
study from Ivory Coast showed a prevalence of 2.8% NHL in adults, consisting of 1.6%
visceral NHL and 1.2% cerebral NHL33. Cerebral lymphoma was reported in only 1%
of cases19,26,33. Only one recent study from South Africa reported heart conditions,
including dilated cardiomyopathy and myocarditis, in 10% of autopsies31. None of the
studies diagnosed HIV-associated multi-centric Castleman disease (MCCD) which is
increasingly recognised in rich countries as an important cause of multi-organ failure
and death in HIV disease34.
HIV 2 infection
Post-mortem findings in HIV-2 infected and dual-HIV reactive patients were similar to
HIV-1 infected patients; only severe multi-organ CMV infection, multinucleate giant
cell encephalitis and cholangitis (secondary to CMV and/or cryptosporidiosis
infection) were significantly more frequent in HIV-2 infected patients, occurring in
18% of HIV-2 cases12,19.
HIV negative adults
Three complete autopsy series and two partial autopsy series reported on autopsy
findings in HIV negative adults (see table 1). The 2 main pathology findings were
pneumonia (22-40%) and malignancies (13-64%). TB was found in 4-16% and caused
death in 4-13%. Cardiovascular diseases were described in 13-27% of cases12,19,23,26.
Abouya et al. report remarkable high prevalence of lung malignancies (64%); 36 %
primary lung carcinoma, 24% metastatic lung cancer and 4% NHL.
Maternal death
HIV-prevalence among 123 women in Mozambique who died during or shortly after
pregnancy in the capital referral hospital was 53%22. In 65 HIV positive women an
autopsy was performed. Seventy one percent of them died of non-obstetric conditions,
including HIV/AIDS related conditions (28%) (15% TB, 8% PJP, 3% KS, 2% malignant
lymphoma), bacterial pneumonia (12%), severe malaria (9%) and bacterial
meningitis (9%). The main obstetric causes of death were puerperal septicaemia
(13%) and eclampsia (13%)22.
Fifty-five percent of the 58 HIV negative women died of non-obstetric conditions,
mainly infections: bacterial pneumonia (16%), severe malaria (14%) and bacterial
meningitis (5%). There were no significant differences in obstetric and non-obstetric
causes of death (HIV related conditions excluded) when comparing HIV-positive to
HIV-negative women22.

32
Pathology findings Children
HIV-positive children
In the 3 complete autopsy series, respiratory tract infections, including TB and PJP
were the main diagnoses (89-94%) and the main cause of death (64-79%)14,20,29.
Bacterial pneumonia was found in 11-58% of children and TB in 1-11%. Interstitial
pneumonitis/bronchiolitis ascertained 6-60% of the pulmonary infections and was
caused by viral infections (CMV, respiratory syncytial virus and adenovirus
infections), HIV-related lymphoid interstitial pneumonia or by an unspecified
organism. Two of the 3 complete autopsy series found PJP to be the main
opportunistic infection in children, with prevalences of 14-28%14,20. PJP was only
found in children under the age of 1 year. The third study did not find any PJP in the
33 children examined, but included only children on cotrimoxazole prophylaxis29.
Disseminated CMV infection was reported in 3 - 43% of children14,20,29.
Enteritis/enterocolitis was reported in approximately 30% of children and caused
death in 0-15%14,20.
Partial and minimal invasive autopsy series documented PJP in 16% of unselected HIV
positive children and in 20-44% of HIV positive children with pulmonary
complaints15-17,27,28,30. PJP was significantly more prevalent in children under 1 year
old and within this group it was more prevalent in the youngest children15,16,27. Other
common pulmonary pathogens were CMV (present in 7-44%) and bacteria (present in
5-41%). TB was found in 5-18% of the specimens obtained by partial autopsies and in
0-4% of the specimens obtained by percutaneous biopsies15-17,27,30.
HIV-negative children
Two complete autopsy series and two partial autopsy series reported on HIV negative
children (table 1)14,15,17,20. Respiratory tract diseases were the dominant causes of
death in all children, but were more often seen in HIV infected children (odds ratio
5.0, 95% confidence interval 2.4-10.5)20. Severe malnutrition was found significantly
more often in HIV positive children and malaria was significantly more often the cause
of death in HIV negative children20. TB was found in 3-8% of the HIV negative
children14,20. PJP was found significantly more often in HIV positive children but was
also found in the lung tissue of 7% of the 84 HIV negative children15.
Correlation clinical and pathology findings
Seven studies compared autopsy and clinical diagnoses (table 2)11,14,21,23-26. The
clinical diagnosis of TB in adults had a sensitivity of 43-80%, a specificity of 67-76%, a
positive predictive value (PPV) of 56%-80% and a negative predictive value (NPV) of
28%11,21,23,24,26. Ordi et al. report on the clinico-pathological discrepancies in diagnosis
of maternal death for the HIV positive and HIV negative autopsy cases together.
Diagnostic errors involving cause of death were detected in 40% of maternal deaths.
The clinical diagnosis of HIV related conditions showed a sensitivity of 33% and a PPV
of 60%25. A clinical diagnosis of PJP in children reached a sensitivity of 50%, a

33
specificity of 91% and a PPV of 71% if a combination of clinical criteria was being
used. Adding chest x-ray results did not improve the accuracy of the clinical
diagnosis14.
Discussion
Major autopsy findings in Sub-Saharan Africa
Autopsy studies show that infectious diseases are almost exclusively responsible for
the pathology in HIV positive individuals in sub-Saharan Africa. TB is the main cause
of death, responsible for almost half of all mortality. PJP is uncommon in adults, but
frequent in children. CMV infection is more common in children. Apart from Kaposi’s
sarcoma, malignancies are uncommonly reported.
None of the studies mentioned that patients had been on ART. With the ongoing
rollout of ART, and with the prolonged use of ART, pathology will probably change19.
Post-mortem studies from developed countries in patients on ART and co-trimoxazole
prophylaxis report a decrease in prevalence of opportunistic infections, a decrease in
Kaposi’s sarcoma, an increase of bacterial infections, and a variable effect on
lymphoma35-37.
In sub-Saharan Africa, so far the only information on causes of death of patients on
ART is based on clinical information from observational cohort studies. These studies
mainly describe mortality within the first 12 months after the start of ART4,6,38.
Leading causes appear to be TB and cryptococcal meningitis, both accounting for
approximately 16% of the early mortality after starting ART4,6,38,39. The impact of IRIS
on mortality seems limited but may be more important when related to central
nervous system infections6,7,38,40,41. Still many patients in sub-Saharan Africa lack
timely access to ART and will therefore present to health care facilities with advanced
HIV disease. The data from the reviewed studies, despite being old, is therefore still
valuable since sadly enough for many patients, circumstances have not yet changed.
Quality of reviewed studies
The quality of the reviewed studies varies widely. Most studies did not describe in
detail their inclusion- and selection procedure. Many studies included less than 50%
of the potential autopsy cases or did not report the number of deaths they did not
include. Information of the cases that were not included regarding e.g. demographics
or clinical diagnosis is often missing. Some studies mention limited resources
compelling them to select only a limited number of patients. Moreover, many studies
did not have microbiology of immunohistochemistry at their disposal. This might have
led to over- or underreporting of some diseases e.g. atypical mycobacterial disease
and lymphoma. Almost no study describes the definitions used to establish clinical or
pathological diagnoses. Many studies did not have access to clinical information to
complete the pathology findings, to establish a final diagnosis.

34
Different studies had different objectives and included different populations, making
it difficult to compare results. Nevertheless, the reviewed studies invariably show the
importance of infectious diseases in all study groups. The comparison of autopsy
findings in HIV positive and HIV negative individuals is difficult since all studies
included HIV negative individuals from wards where the HIV prevalence was high and
therefore not representative of all HIV negative deaths.
Correlation clinical and pathology findings
The described autopsy series show a weak correlation between clinical diagnosis and
pathology findings. Possible reasons can be a low level of suspicion by clinicians,
atypical disease-presentation, a lack of resources to perform investigations, poor yield
of diagnostics and/or a selection bias in the reviewed studies with over-
representation of patients with an unknown clinical cause of death. Another problem
is the high proportion of patients admitted to the hospital with first time HIV
diagnosis and advanced clinical disease. These patients are often either too sick to
perform diagnostic tests and/or die before any diagnostic tests can be performed.
Finally, many studies were done 10-20 years ago, when the clinical learning curve of
HIV-related clinical pathology was still upward.
Value of autopsy
Autopsy is an important tool. For the bereaved relatives it can answer questions
regarding the death of a beloved one and reveal conditions that might have
consequences for family members in the longer term. On an institutional level,
autopsies can be used for teaching purposes and quality control of the clinical care.
Two examples of how pathology findings can influence clinical practise come from the
Ivory Coast autopsy cohort19,32,33. This study showed the ratio of central nervous
system (CNS) toxoplasmosis to lymphoma was so high that empirical toxoplasmosis
treatment for focal CNS lesions was the rational approach12. Also the 4% prevalence of
nocardiosis, an imitator of TB both clinically and radiologically, indicated – yet again –
that the diagnosis and management of (smear negative) TB is not always
straightforward (Lucas, S., personal communication).
For governments, widespread autopsy practices are the ultimate tool to provide
epidemiological data on changing and emerging disease patterns, particularly relating
to infectious diseases. Reliable information should guide resource allocation. The
pathophysiology of relative new phenomena, like IRIS, can be studied. Clinically
unsuspected conditions including drug toxicities can be revealed and studied at a
cellular level. And it is likely that unanticipated new syndromes will emerge among
long-term ART-treated patients, which may be drug-related or novel HIV-immune-
related; autopsies should identify these.

35
Challenges to perform autopsies
As the reviewed studies show, autopsies can be done in sub-Saharan Africa. However,
a number of obstacles impede its widespread use. To perform autopsies you need a
well-equipped mortuary and histopathology department with trained staff able to
perform autopsies in a safe manner. Also, ready access to microbiology diagnostics
should be available. These requirements are costly and within Sub-Saharan Africa
often only met in a very limited number of large tertiary teaching facilities or large
international study settings. So feasibility of autopsy as a common-practice
intervention is limited and autopsy study results are often only applicable for patients
hospitalised in large tertiary teaching facilities. Moreover, standard guidelines for
performing autopsies and evaluating the histological findings are lacking.
Another difficulty that arises at population level is the poor cultural and social
acceptance of autopsy practices. This negatively influences both request- and
acceptance rates. There are at least four autopsy scenarios: consent given by relatives;
medico-legal instruction for autopsy without consent (i.e. forensic case); autopsy by
default with opt-outs by relatives (none in this study set); and no consent required at
all and no opt-outs (as pertained in four studies12,19,20,29). Only the last scenario can
provide an epidemiologically representative sampling of cases. However, it is
questionable how ethical this scenario is, since it disregards the opinion and emotions
of relatives and patients.
As shown by Menendez et al., autopsy rates of 78% can be reached, with consent from
family members and without incentive. This is in contrast with the belief that within
African society, autopsies are unacceptable because of cultural or religious reason.
Two studies from sub-Saharan Africa report on factors interfering with acceptance of
autopsies by relatives. They found that the uselessness of the results for the relatives,
fear of mutilation, concern about delaying the funeral or prior transport arrangements
made, traditional beliefs and objection by the deceased prior to death negatively
influenced the acceptance of autopsy by the relatives42,43.
Alternative methods
An option to solve some of the feasibility and acceptance issues is to perform partial
or minimal invasive autopsies. In the studies we reviewed, those that performed
partial and minimal invasive autopsies included 25-65% of the deaths, compared to 5-
78% in those that performed complete autopsies. Because partial and minimal
invasive autopsies can focus on an organ of particular interest, they take less time, are
less costly and less mutilating. The trade-off is incomplete information due to
sampling error and the loss of visual judgement of all organs by the pathologist. A
study comparing results of needle autopsies with conventional autopsies in adults
found a concordance in the cause of death of 67%. In the same study, however,
conventional autopsies found an additional major diagnosis in 87% of cases44. In

36
neonates full concordance in cause of death between conventional autopsy and needle
autopsy results was found in 56% of cases and partial concordance in another 12%45.
Non-invasive autopsy practices, using magnetic resonance imaging (MRI) and/or
multi-slice computer tomography (MSCT), can be of additional value if combined with
minimally invasive autopsy, but are not yet well enough validated to be used as a
replacement for the conventional autopsy46,47. Considering the high costs of these
techniques and the low availability they are however not considered a suitable
alternative for sub-Saharan Africa in the near future.
Verbal autopsy procedures are an indirect retrospective method to identify the
probable cause of death, by questioning the relatives or other associates of the
deceased. Afterwards a doctor reviews the answers and establishes a cause of death.
This relatively cheap method is used mainly in resource-limited settings to acquire
knowledge on causes of mortality48. The information gathered relies on clinical record
keeping and the (long-term) memory of relatives or other associates of the deceased.
So far, verbal autopsies have only been validated with clinical diagnoses as gold
standard49,50. Given the earlier discussed high discordance between clinical and
postmortem diagnoses, the accuracy and therefore the value of verbal autopsies for
HIV-infected persons is very questionable.
Conclusion
Autopsy is an important tool to develop more evidenced-based diagnostic and
treatment guidelines, to detect and document adverse events of new treatments, to
improve our understanding of the pathophysiology of HIV infection, to discover future
research areas, to provide epidemiologic data on disease patterns and to detect
medical errors and weaknesses in health care delivery.
Autopsy studies are possible in Africa, yet hurdles need to be overcome to allow a
more widespread practice. So far, nearly all autopsy studies in sub-Saharan Africa
were performed before the introduction of ART and none of them included patients
on ART. In view of the persistent high mortality rates in sub-Saharan Africa and the
large-scale use of new treatment regimens, it remains important to determine causes
of death, identify emerging syndromes, and describe drug-induced pathology. Partial
autopsies can adequately address certain clinical and research questions but complete
autopsies should remain the gold standard in individual patient care. Verbal
autopsies, although a feasible alternative, still need adequate validation before solid
conclusions can be drawn from their results.

Table 1. Study and patient characteristics and main autopsy findings in HIV positive, and *HIV negative individuals
First author,
Country,
Publication yr
Type of study, Setting,
Population, Specification
autopsy*
Nmb of
autopsies/n
mb of deaths
Median
age
Sex:
% ♂
Main causes of death Main pathology findings
COMPLETE AUTOPSIES ADULTS
Lucas,
Ivory Coast,
1993 19
Prospective,
2 large urban hospitals,
different medical wards +
community deaths
Age >14 years, known
serostatus, preference HIV-2
247/1020
(24%)
36 65 Tuberculosis 32%
Bacteriaemia 11%
Cerebral toxoplasmosis 10%
Bacterial pneumonia 8%
Bacterial meningitis 5%
Tuberculosis 38%
CMV 18%
Bacteriaemia 16%
Non Hodgkin Lymphoma 3%
Lucas,
Ivory Coast,
1993 19
HIV negative individuals 42/518 (8%) ♂ 43
♀ 22
69 Bacterial pneumonia 22%
Cancer 13%
Hypertension 13%
Tuberculosis 7%
Rana, Kenya,
1999 26
Prospective,
2 acute medical wards of a
tertiary hospital
Adults, known serostatus
75/155
(48%)
Interstitial pneumonia 15%
Cryptococcal disease 4%
CMV pneumonitis 4%
Tuberculosis 50%
Lymphocytic meningitis 11%
(Multi)focal bacterial sepsis 11%
Rana, Kenya, 1999
26
HIV negative individuals 47/141
(16%)
35 57 Bacterial pneumonia 38%
Tuberculosis 13%
Malignancy 9%
Malaria 6%
Malignancy 26%
Tuberculosis 15%
Sepsis 9%
Ansari, Botswana,
2002 13
Prospective,
All medical wards of a referral
hospital
Age >13 years, mainly with
pulmonary symptoms
104/565
(18%)
Pneumonia 14%
PJP 11%
Kaposi’s sarcoma 7%
Cryptococcosis 6%
Tuberculosis 40%
Pneumonia 23%
CMV 15%
PJP 11%
Ansari, Botswana,
2002 13
HIV negative individuals 24/380 (6%) 53 54 Pneumonia 21%
Cardiac disease 21%
Pneumonia 29%
Cardiac disease 17%

38
Non Hodgkin lymphoma 13%
Cerebral haemorrhage 5%
Intestinal infarction 5%
Other malignancies 5%
Cerebral haemorrhage 5%
Intestinal infarction 5%
Other malignancies 5%
PARTIAL AUTOPSIES ADULTS
Abouya, Ivory
Coast, 1992 12
Prospective,
Pulmonary ward of university
hospital,
Age >14 years, known
serostatus,
Lungs only
53/71 (75%) - - Pulmonary TB 40%
Non-specific pneumonia 34%
PJP 8%
Other cancers 4%
Pulmonary tuberculosis 43%
Non-specific pneumonia 34%
PJP 9%
Other cancers 4%
Abouya, Ivory
Coast, 1992 12
HIV negative individuals 25/29 (86%) - - Cancer, primary or secondary
64%
Pneumonia n.s. 28%
Cancer, primary or secondary 64%
Pneumonia n.s. 28%
Tuberculosis 12%
Nelson, Zaire,
1993 24
Prospective,
3 hospitals: urban general,
urban referral, rural
Adults, AIDS on chart or death
certificate, death shortly after
admission of unknown cause,
Complete, except brain
63/- 36 44 - Extrapulmonary TB 41%
Candidiasis 31%
Extrapulmonary cryptococcosis
19%
CMV 13%
Kibayashi,
Tanzania, 1999 18
Prospective,
1615 forensic autopsies, 30
HIV+ suspects selected, 10
confirmed HIV+
University hospital
Focus on neuropathology
Sampling of brain, heart, lungs,
liver and kidneys
10/- 29 100 - Neuropathology findings
Lymphocytic meningitis 50%
TB meningitis + abces 10%
Cryptococcal meningitis 10%
Intracerebral haemorrhage 10%
No abnormalities 20%
Agyei, Ghana,
2004 11
Prospective,
HIV unit in major teaching
hospital,
Adults, HIV +,
Autopsy procedure n.s.
134/ 224
(60%)
- - - Tuberculosis 54%
Malignancies 6%
Murray, South Prospective cohort of miners** 66/242 - 100 - Respiratory infection n.s 83%

39
Africa, 2007 23 Cardio-respiratory organs (27%) Pulmonary tuberculosis 21%
Cryptococcal pneumonia 17%
PJP 14%
Murray, South
Africa, 2007 23
(33%)
- 100 - Any respiratory infection 37%
TB STUDIES COMPLETE AUTOPSY
Martinson, South
Africa, 2007 21
Prospective,
2 sites: large tertiary hospital
and inpatient TB facility
TB suspects, > 18 years
50/1000
(5%)***
(47 HIV+)
Interstitial pulmonary disease
5%
CMV pneumonitis 4%
PJP 4%
Tuberculosis 79%
CMV pneumonitis 11%
PJP 6%
MATERNAL MORTALITY COMPLETE AUTOPSIES
Menendez,
Mozambique,
2008 22
Prospective,
Tertiary hospital,
Women who died during
pregnancy or within 42 days
after delivery
139/179
(78%)***
(65 HIV+)
- - Non-obstetric conditions 71%
HIV-related conditions 28%:
Tuberculosis 15%
PJP 8%
Malignant lymphoma 2%
-
Menendez,
Mozambique,
2008 22
HIV negative individuals 58/- 0 Non obstetric conditions 55%:
Bacterial pneumonia 16%,
Severe malaria 14%,
Neoplasm 5%
Other and unknown 10%
-
CHILDREN COMPLETE AUTOPSY
Lucas, Ivory
Coast, 1996 20
Prospective,
Mortuary based, largest
hospital of Abidjan
+ community deaths
Age 1 month-12 yrs
78/80 (98%) 18 months 50 Pneumonia 41%
Measles 17%
Meningitis 15%
PJP 11%
Enteritis 8%
Severe malnutrition 55%
CMV infection 32%
Enteritis 30%
Interstitial
pneumonitis/bronchiolitis 18%
Lucas, Ivory Coast,
1996 20
(23%)
21 months - Malaria 23%
Pneumonia (PJP excluded) 22%
Meningitis 12%
Measles 12%
Malaria 30%
Severe malnutrition 26%
Enteritis 18%

40
Other 31% Interstitial
Chakraborty,
Kenya, 2002 29
Retrospective,
Urban orphanage ,
Age 1 month-18 years
33/33
(100%)
71 months 39 Bacterial pneumonia 58%
Herpes encephalitis 6%
Disseminated TB 3%
Dilated cardiomyopathy 21%
Myocarditis 12%
Nephritis 12%
Ansari, Botswana,
2003 14
Prospective,
Referral hospital,
Age 1 month-13 years
35/126
(28%)
7 months 45 Interstitial
PJP 29%
Pneumonia 11%
Tuberculosis 11%
Disseminated CMV 11%
Interstitial
Disseminated CMV 43%
Enterocolitis 31%
PJP 29%
HIV encephalitis 14%
Ansari, Botswana,
2003 14
(10%)
16 months - Pneumonia 42%
Interstitial
Cardiomyopathy/myo-or
pericarditis 17%
Tuberculosis 8%
Enterocolitis 50%
Pneumonia 42%
Cardiomyopathy/myo-or
pericarditis 25%
Interstitial
Tuberculosis 8%
CHILDREN PARTIAL/ MINIMAL INVASIVE AUTOPSIES
Ikeogu,
Zimbabwe, 1997
16
Prospective,
Referral and teaching hospital,
Age <5 years, <3 hours dead on
arrival or dead shortly after
arrival,
Lung sampling
122/ 334
(37%)
10 months 51 - PJP 16%
Lymphoid interstitial pneumonia
9%
CMV 7%
Tuberculosis 5%
Jeena, South
Africa, 1996 17
Prospective,
ICU unit of large hospital,
Percutaneous biopsies of lung
and liver
31/43 lung
(72%)
34/43 liver
(79%)
4,3
months
45 - Lung biopsies:
PJP 52%
CMV 52%
Liver biopsies:
Steatosis 65%
CMV 15%
Jeena, South
Africa, 1996 17
HIV negative individuals 34/**** 4 months 45 - Lung biopsies:

41
Unspecified pneumonia 18%
Adenovirus 15%
Liver biopsies;
Steatosis 59%
CMV 6%
Nathoo,
Zimbabwe, 2001
30
Prospective,
University hospital
Children who died of
pneumonia
4 percutaneous biopsies of lung
21/618 (3%) 3 months 46 PJP 29%
CMV 10%
Chintu, Zambia,
2002 15
Prospective
Referral and teaching hospital,
Age 1 month-16 years admitted
with respiratory illness
Chest only
264/ 1603
(18%)***
180 HIV+
8 months 52 - Bacterial pneumonia 41%
PJP 29%
CMV 22%
Tuberculosis 18%
Shock lung 13%
Chintu, Zambia,
2002 15
HIV negative individuals 84/- 11,5 52 - Bacterial pneumonia 50%
Tuberculosis 26%
Interstitial pneumonitis 18%
Pulmonary oedema 11%
PJP 7%
Rennert, South
Africa, 2002 27
Prospective,
Large teaching hospital,
premortem diagnosis of lung
disease,
Percutaneous biopsies of lung
and liver
93/- 10.5
months
53 CMV 31%
PJP 20%
Viral infection other than CMV 14%
Pneumonitis n.s. 12%
Ruffini, South
Africa, 2002 28
Prospective,
Secondary and tertiary level
hospital,
Age <2 years,
HIV+ or clinically suspected
HIV+
admitted with severe
pneumonia,
Percutaneous lung biopsies
18/ 29
(62%)
3.5
months
- PJP 44%
CMV pneumonia 44%
Non-specific interstitial
pneumonitis 28%
Complete autopsies combined adults, maternal mortality and children

42
Garcia-Jardon,
South Africa,
2010 31
Retrospective 2000-2005
Prospective 2006-2008
Tertiary level hospital
All HIV+ deaths
86/- 58% 18-
55 yrs
36% < 18
yrs
35 Disseminated TB 19%
Pulmonary TB 13%
Pneumonia 22%
Liver failure 10%
Heart conditions 10%
CMV = cytomegalo virus; PJP Pneumocystis jiroveci pneumonia; n.s. not specified *
When applicable ** We report only on those that died of a natural cause *** Autopsies HIV+ and HIV - / total number of death **** Age, sex
and race matched controls out of 123 autopsies in HIV- children

43
Table 2. Comparison of clinical versus autopsy diagnosis
First author Clinical
case
definition
Source of clinical
diagnosis
Diagnosis Comparison
clinical/
autopsy
diagnosis
ADULTS
Agyei 11 Not defined Not defined Tuberculosis Sens 59%
Murray 23 Not defined Mine personnel records,
assurance records,
hospital records,
Employment Bureau
records, national death
register, autopsy
database at National
Institute of Occupational
Health
Pulmonary
tuberculosis
Respiratory
infection
Sens 43%
Spec 67%
Sens: 51%
Spec 55%
Nelson 24 Not defined Chart review Tuberculosis
Overall
diagnosis
PPV 56%
Sens 27%
Rana 26 Not defined Recorded at death by
study team
Tuberculosis Sens 54%
Spec 74%
TUBERCULOSIS SUSPECTS
Martinson 21 Not defined Pre-mortem diagnosis by
attending doctor
Tuberculosis PPV 79-80%
NPV 28%
MATERNAL MORTALITY
Ordi 25 Not defined Listing by the clinician HIV related
conditions
Sens 33%
Spec 97%
PPV 67%
NPV 90%
CHILDREN
Ansari 14 Using
several
case-
definitions
Medical records Pneumocystis
jiroveci
pneumonia
Sens 40-100%
Spec 48-91%
PPV 36-71%

44
References
1. Reniers G, Araya T, Davey G, et al. Steep declines in population-level AIDS mortality
following the introduction of antiretroviral therapy in Addis Ababa, Ethiopia. AIDS. Feb 20
2009;23(4):511-518.
2. Brinkhof MW, Boulle A, Weigel R, et al. Mortality of HIV-infected patients starting
antiretroviral therapy in sub-Saharan Africa: comparison with HIV-unrelated mortality. PLoS
Med. Apr 28 2009;6(4):e1000066.
3. UNAIDS WHOa. AIDS Epidemic Update. Geneva: World Health Organisation and
UNAIDS;December 2009.
4. Lawn SD, Harries AD, Anglaret X, Myer L, Wood R. Early mortality among adults
accessing antiretroviral treatment programmes in sub-Saharan Africa. AIDS. Oct 1
2008;22(15):1897-1908.
5. Braitstein P, Brinkhof MW, Dabis F, et al. Mortality of HIV-1-infected patients in the
first year of antiretroviral therapy: comparison between low-income and high-income
countries. Lancet. Mar 11 2006;367(9513):817-824.
6. Castelnuovo B, Manabe YC, Kiragga A, Kamya M, Easterbrook P, Kambugu A. Cause-
specific mortality and the contribution of immune reconstitution inflammatory syndrome in the
first 3 years after antiretroviral therapy initiation in an urban African cohort. Clin Infect Dis. Sep
15 2009;49(6):965-972.
7. Lawn SD, Myer L, Orrell C, Bekker LG, Wood R. Early mortality among adults accessing
a community-based antiretroviral service in South Africa: implications for programme design.
AIDS. Dec 2 2005;19(18):2141-2148.
8. MacPherson P, Moshabela M, Martinson N, Pronyk P. Mortality and loss to follow-up
among HAART initiators in rural South Africa. Trans R Soc Trop Med Hyg. Jun 2009;103(6):588-
593.
9. Cooper M, van der Loeff MS, McConkey S, Cooper S, Vincent T, Whittle H. Cause of
death among people with retroviral infection in rural Guinea-Bissau. Trop Doct. Jul;40(3):181-
183.
10. Beadsworth MB, Cohen D, Ratcliffe L, et al. Autopsies in HIV: still identifying missed
diagnoses. Int J STD AIDS. Feb 2009;20(2):84-86.
11. A. Agyei ML. Spectrum of Opportunistic Infections Causing Death in Patients with AIDS
in Ghana: Correlation of Clinical Diagnoses with Autopsy Results. XI Conference on Retrovirus
and Opportunistic Infections; 8-11 February, 2004; San Francisco.
12. Abouya YL, Beaumel A, Lucas S, et al. Pneumocystis carinii pneumonia. An uncommon
cause of death in African patients with acquired immunodeficiency syndrome. Am Rev Respir
Dis. Mar 1992;145(3):617-620.
13. Ansari NA, Kombe AH, Kenyon TA, et al. Pathology and causes of death in a group of
128 predominantly HIV-positive patients in Botswana, 1997-1998. Int J Tuberc Lung Dis. Jan
2002;6(1):55-63.
14. Ansari NA, Kombe AH, Kenyon TA, et al. Pathology and causes of death in a series of
human immunodeficiency virus-positive and -negative pediatric referral hospital admissions in
Botswana. Pediatr Infect Dis J. Jan 2003;22(1):43-47.
15. Chintu C, Mudenda V, Lucas S, et al. Lung diseases at necropsy in African children
dying from respiratory illnesses: a descriptive necropsy study. Lancet. Sep 28
2002;360(9338):985-990.
16. Ikeogu MO, Wolf B, Mathe S. Pulmonary manifestations in HIV seropositivity and
malnutrition in Zimbabwe. Arch Dis Child. Feb 1997;76(2):124-128.
17. Jeena PM, Coovadia HM, Chrystal V. Pneumocystis carinii and cytomegalovirus
infections in severely ill, HIV-infected African infants. Ann Trop Paediatr. Dec 1996;16(4):361-
368.
18. Kibayashi K, Ng'walali PM, Mbonde MP, et al. Neuropathology of human
immunodeficiency virus 1 infection. Significance of studying in forensic autopsy cases at Dar es
Salaam, Tanzania. Arch Pathol Lab Med. Jun 1999;123(6):519-523.
19. Lucas SB, Hounnou A, Peacock C, et al. The mortality and pathology of HIV infection in
a west African city. AIDS. Dec 1993;7(12):1569-1579.

45
20. Lucas SB, Peacock CS, Hounnou A, et al. Disease in children infected with HIV in
Abidjan, Cote d'Ivoire. BMJ. Feb 10 1996;312(7027):335-338.
21. Martinson NA, Karstaedt A, Venter WD, et al. Causes of death in hospitalized adults
with a premortem diagnosis of tuberculosis: an autopsy study. AIDS. Oct 1 2007;21(15):2043-
2050.
22. Menendez C, Romagosa C, Ismail MR, et al. An autopsy study of maternal mortality in
Mozambique: the contribution of infectious diseases. PLoS Med. Feb 2008;5(2):e44.
23. Murray J, Sonnenberg P, Nelson G, Bester A, Shearer S, Glynn JR. Cause of death and
presence of respiratory disease at autopsy in an HIV-1 seroconversion cohort of southern
African gold miners. AIDS. Nov 2007;21 Suppl 6:S97-S104.
24. Nelson AM, Perriens JH, Kapita B, et al. A clinical and pathological comparison of the
WHO and CDC case definitions for AIDS in Kinshasa, Zaire: is passive surveillance valid? AIDS.
Sep 1993;7(9):1241-1245.
25. Ordi J, Ismail MR, Carrilho C, et al. Clinico-pathological discrepancies in the diagnosis
of causes of maternal death in sub-Saharan Africa: retrospective analysis. PLoS Med. Feb 24
2009;6(2):e1000036.
26. Rana FS, Hawken MP, Mwachari C, et al. Autopsy study of HIV-1-positive and HIV-1-
negative adult medical patients in Nairobi, Kenya. J Acquir Immune Defic Syndr. May 1
2000;24(1):23-29.
27. Rennert WP, Kilner D, Hale M, Stevens G, Stevens W, Crewe-Brown H. Tuberculosis in
children dying with HIV-related lung disease: clinical-pathological correlations. Int J Tuberc
Lung Dis. Sep 2002;6(9):806-813.
28. Ruffini DD, Madhi SA. The high burden of Pneumocystis carinii pneumonia in African
HIV-1-infected children hospitalized for severe pneumonia. AIDS. Jan 4 2002;16(1):105-112.
29. Chakraborty R, Pulver A, Pulver LS, et al. The post-mortem pathology of HIV-1-infected
African children. Ann Trop Paediatr. Jun 2002;22(2):125-131.
30. Nathoo KJ, Gondo M, Gwanzura L, Mhlanga BR, Mavetera T, Mason PR. Fatal
Pneumocystis carinii pneumonia in HIV-seropositive infants in Harare, Zimbabwe. Trans R Soc
Trop Med Hyg. Jan-Feb 2001;95(1):37-39.
31. Garcia-Jardon M, Bhat VG, Blanco-Blanco E, Stepian A. Postmortem findings in
HIV/AIDS patients in a tertiary care hospital in rural South Africa. Trop Doct. Apr;40(2):81-84.
32. Lucas SB, Hounnou A, Peacock C, Beaumel A, Kadio A, De Cock KM. Nocardiosis in HIV-
positive patients: an autopsy study in West Africa. Tuber Lung Dis. Aug 1994;75(4):301-307.
33. Lucas SB, Diomande M, Hounnou A, et al. HIV-associated lymphoma in Africa: an
autopsy study in Cote d'Ivoire. Int J Cancer. Oct 1 1994;59(1):20-24.
34. Oksenhendler E. HIV-associated multicentric Castleman disease. Curr Opin HIV AIDS.
Jan 2009;4(1):16-21.
35. Lyon R, Haque AK, Asmuth DM, Woods GL. Changing patterns of infections in patients
with AIDS: a study of 279 autopsies of prison inmates and nonincarcerated patients at a
university hospital in eastern Texas, 1984-1993. Clin Infect Dis. Aug 1996;23(2):241-247.
36. Vago L, Bonetto S, Nebuloni M, et al. Pathological findings in the central nervous
system of AIDS patients on assumed antiretroviral therapeutic regimens: retrospective study of
1597 autopsies. AIDS. Sep 27 2002;16(14):1925-1928.
37. Masliah E, DeTeresa RM, Mallory ME, Hansen LA. Changes in pathological findings at
autopsy in AIDS cases for the last 15 years. AIDS. Jan 7 2000;14(1):69-74.
38. Etard JF, Ndiaye I, Thierry-Mieg M, et al. Mortality and causes of death in adults
receiving highly active antiretroviral therapy in Senegal: a 7-year cohort study. AIDS. May 12
2006;20(8):1181-1189.
39. Munderi P. WC, Nakiyingi J., Kasirye A., Walker S., French N., Gilks C., Grosskurth H.
Survival and Causes of Death, 2 years after introduction of Antiretroviral Therapy in Africa: a
historical cohort comparison in Entebbe, Uganda. Paper presented at: International AIDS
conference 13 - 18 August, 2006; Toronto, Canada.
40. Bicanic T, Meintjes G, Rebe K, et al. Immune reconstitution inflammatory syndrome in
HIV-associated cryptococcal meningitis: a prospective study. J Acquir Immune Defic Syndr. Jun 1
2009;51(2):130-134.

46
41. Lawn SD BL, Myer L, Orrell C., Wood R. Cryptococcocal immune reconstitution disease:
a major cause of early mortality in a South African antiretroviral programme. . AIDS.
2005;19:2050-2052.
42. Oluwasola OA, Fawole OI, Otegbayo AJ, Ogun GO, Adebamowo CA, Bamigboye AE. The
autopsy: knowledge, attitude, and perceptions of doctors and relatives of the deceased. Arch
Pathol Lab Med. Jan 2009;133(1):78-82.
43. Lishimpi K, Chintu C, Lucas S, et al. Necropsies in African children: consent dilemmas
for parents and guardians. Arch Dis Child. Jun 2001;84(6):463-467.
44. Huston BM, Malouf NN, Azar HA. Percutaneous needle autopsy sampling. Mod Pathol.
Dec 1996;9(12):1101-1107.
45. Garg S, Punia RP, Basu S, Mohan H, Bal A. Comparison of needle autopsy with
conventional autopsy in neonates. Fetal Pediatr Pathol. 2009;28(3):139-150.
46. Burton JL, Underwood J. Clinical, educational, and epidemiological value of autopsy.
Lancet. Apr 28 2007;369(9571):1471-1480.
47. Roberts ISD, Benbow EW, Bisset R, et al. Accuracy of magnetic resonance imaging in
determining cause of sudden death in adults: comparison with conventional autopsy.
Histopathology. 2003;42(5):424-430.
48. WHO. Report of the WHO technical consultation of verbal autopsy tools. Review of the
literature and currently used verbal autopsy tools. Talloires, France2004.
49. Freeman JV, Christian P, Khatry SK, et al. Evaluation of neonatal verbal autopsy using
physician review versus algorithm-based cause-of-death assignment in rural Nepal. Paediatr
Perinat Epidemiol. Jul 2005;19(4):323-331.
50. Lopman B, Cook A, Smith J, et al. Verbal autopsy can consistently measure AIDS
mortality: A validation study in Tanzania and Zimbabwe. J Epidemiol Community Health. Oct 23
2009.

47
Chapter 3
Temporal Trends in All Cause Mortality in Adults
attending an Urban HIV-clinic in Uganda: a
Retrospective Chart-review
Janneke A. Cox1
, Daniel Kiggundu2
, Lana Elpert3
, Graeme Meintjes4
, Robert
Colebunders1,5
, Stella Alamo6
1
Department of Clinical Sciences, Institute of Tropical Medicine, Antwerp, Belgium
2
Department of Medicine, Makerere University College of Health Sciences, Kampala,
Uganda
3
Department of Medicine, Johns Hopkins University School of Medicine, Baltimore,
Maryland, United States of America
4
Institute of Infectious Disease and Department of Medicine, University of Cape Town,
Cape Town, South Africa
5
Faculty of Medicine, University of Antwerp, Belgium
6
Reach Out Mbuya Parish HIV/AIDS Initiative, Kampala, Uganda
Submitted

49
Abstract
Objective
A shift from HIV-related causes of death towards non HIV-related causes will impact
patient care and resource allocation. We studied all-cause mortality in HIV-infected
adults that attended an HIV-clinic in Kampala, Uganda, between 2001 and 2012 to see
if any temporal trends in causes of death could be observed.
Methods
Two doctors independently reviewed the clinic database that contained information
derived retrospectively from the clinic files and assigned one or more causes of death
to each patient >18 years of age with a known date of death. Four cause-of-death
categories were defined: ‘communicable conditions and AIDS-defining malignancies’,
‘chronic non-communicable conditions’, ‘other non-communicable conditions’ and
‘unknown’. Trends in cause-of-death categories over time were evaluated using
logistic regression with year of death as an independent variable.
Results
1028 deaths were included; 38% were on ART. Over time, the proportion of deaths
from ‘communicable conditions and AIDS-defining malignancies’ significantly
decreased. The proportion of deaths from ‘chronic non-communicable conditions’,
‘other non-communicable conditions’ and a combination of ‘communicable and non-
communicable conditions’ significantly increased. The main cause of death was TB
(34%). Causes of death in those on ART > 1 year (n=127) were unspecified infection
(18%), TB (14%), accident, suicide and assault (10%), malignancies (9%) and
meningitis (8%).
Conclusion
We observed a decrease over time in the proportion of deaths from communicable
conditions and AIDS-defining malignancies and an increase in non-communicable
diseases, both chronic and non-chronic. Nevertheless, communicable conditions and
AIDS-defining malignancies continued to cause the majority of deaths, with TB as
main cause. Ongoing monitoring of cause of death is warranted and strategies to
decrease mortality from TB and other common opportunistic infections are essential.

50
Background
Over the last decade the landscape of HIV-care and treatment in sub-Saharan Africa
(SSA) has changed tremendously. Testing and counseling services have been scaled
up, antiretroviral therapy (ART) has become more widely available and guidelines
suggesting initiating ART at higher CD4 counts have been issued1, 2. In Uganda, the
number of HIV-infected people on ART between 2001 and 2012 increased from
scarcely any to almost 440.0002. Moreover, improved diagnostic tests for common
opportunistic infections like TB are being implemented3, 4. All of these measures have
resulted in a decline in mortality amongst HIV-infected individuals throughout SSA2, 5-
7.
In reports to date, the majority of deaths among HIV-infected adults in SSA have been
secondary to HIV-related infections and malignancies8, 9. However, with longer
survival of HIV-infected individuals, a transition in causes of morbidity and mortality
can be expected10. Changes in socio-economic development, lifestyle and dietary
intake will impact the causes of death in the population as a whole, including HIV-
infected individuals11. ART-related toxicity and (low-level) chronic inflammation play
an additional role in HIV-infected patients. All of these factors may induce a shift
towards causes of death that are not directly related to HIV. In countries where ART
has been available for a longer time, such a shift is already seen and hepatic diseases,
cardiovascular diseases and non-AIDS malignancies have become common causes of
death12-14. Awareness of such a changing pattern in the African setting is of
importance for patient care and resource allocation. Current HIV-care in SSA is largely
focused on management of opportunistic co-infections. If non-infectious chronic
medical conditions become increasingly responsible for death among HIV-infected
individuals, a shift of focus would be needed15.
We evaluated all-cause mortality in HIV-infected adults that attended an urban HIV-
treatment facility in Kampala, Uganda. We retrospectively studied the ante-mortem
characteristics, the clinical course of the deceased and the causes of death since the
start of the clinic in 2001 until 2012 to see if any temporal trends in causes of death
could be observed.
Methods
Study setting and population
We conducted a retrospective study at the Reach Out Mbuya Parish HIV/AIDS
initiative (ROM) in Kampala. ROM’s HIV-treatment program has been described in
detail previously16, 17. In brief, ROM has provided free individual support services and
community-programs to HIV-infected patients living in the urban slums that surround
the parish since May 2001. Free ART first became available in June 2003 through
research based clinics and individual sponsorships. In March 2004 ROM was the first
program to initiate ART through PEPFAR. However, the demand was greater than the

51
supply and ART initiation was based on the principle to serve those with the lowest
CD4 cell count first. ART eligibility was based on the WHO guidelines adopted for use
through the Uganda National ART guidelines and subsequently followed the evolving
recommendations to initiate ART at a higher CD4 cell count.
CD4 count measurement has been available since 2004, but viral load measurement
was never routinely available. Community health care workers (CHW), trained peers
that support each patient, conduct home visits at intervals from daily to once per 3
months based on the needs of the specific patient. Bedridden or hospitalized patients
are visited weekly by CHWs. Any relevant medical information obtained during their
visits is noted in the clinic file. In addition to the home visits, patients visit the clinic at
intervals from every 2 weeks to once per 3 months depending on their treatment
status.
Tracing and registration of patients
Patients who miss their clinical appointment are visited by a CHW. Until October
2008, a handwritten register was used to identify missed appointments, and it could
take up to one month after the missed appointment that a CHW visit was made. Since
then an electronic medical record system has been installed that identifies missed
appointments the same day and allows for home visits by the CHW within 24 hours. In
an earlier evaluation, this system reduced both missed appointments and patients lost
to follow up (LTFU) substantially18.
In the clinic registration system patients are classified as ‘active’, ‘dead’, ‘LTFU’ or
‘transferred out’. A patient is classified as ‘dead’ if the clinic or CHW is informed about
the patients' death. If a patient misses clinic appointments for >90 days, and the CHW
is unable to gather additional information, he/she is classified as ‘LTFU’. Because of
the efficient community network, in most instances, reliable information about a
death can be collected within a week of the patient’s death.
Mortality audit
In March 2013, ROM established a database including all patients that had died since
the start of the clinic in May 2001. A trained administrative assistant extracted the
following information from the clinic file of each patient known to have died: date of
HIV diagnosis; date of registration; CD4 count at registration, at its lowest value, and
the most recent CD4 prior to death; if the patient was started on ART, the date ART
was started and the regimen; the date of death; and the last weight and the height
recorded prior to death. Two experienced medical doctors (LE, DK) then reviewed the
clinic file and summarized the clinically relevant events preceding the death. In case of
any inconsistency in the database, the doctor would review the patients’ clinic file and
adjust the information accordingly.

52
Assigning the cause of death
Two study team doctors (DK, JC) independently reviewed the database. A cause of
death was assigned to each adult patient based on a listing of predefined causes of
death, which was derived from the global burden of disease listing19. Patients below
18 years of age or with an unknown death date were excluded. When no specific cause
of death could be assigned, but the cause of death was considered to be related to a
communicable disease e.g. because of the presence of fever, the category
‘communicable disease unspecified’ was assigned, and the organ system involved
(respiratory, central nervous system, gastro-intestinal, other or unclear) indicated.
This was the same for ‘non-communicable diseases unspecified’. HIV as cause of death
was assigned only to patients with a CD4 count below 100 cells/mm3 in the 6 months
preceding death without the start of ART and no apparent other cause of death. Two
causes of death could be assigned to one patient if 2 diseases were thought to have
contributed equally to a patient’s death (e.g. pulmonary TB and diabetic keto-
acidosis). In cases where no distinction could be made between a communicable or
non-communicable cause of death, the category ‘unknown’ was assigned. The cause-
of-death listings of both doctors were compared, and in cases of a discrepancy, it was
discussed with a third doctor (RC) to assign a final consensus cause of death.
Outcomes and statistical methods
Causes of death were divided into 4 categories: 1) communicable conditions and
AIDS-defining malignancies, 2) chronic non-communicable conditions, 3) other non-
communicable conditions which included pregnancy-related death, external causes
(accidents, suicide, assault) and post-operative death and 4) unknown. To evaluate
any trends in cause of death categories over time, we used logistic regression with
year of death as an independent variable. Each category of cause of death was
separately tested against the total of all other categories. A p-value ≤0.05 was
considered statistically significant. Bonferroni correction for multiple testing was
used. For comparison of non-Normally distributed variables, Wilcoxon rank sum test
was performed. A p-value ≤0.05 was considered statistically significant.
Proportions are reported with a Wilson 95% confidence interval (CI) and medians
with an interquartile range (IQR). Person years of follow up for the whole clinic
population were not available; therefore, mortality ratios were calculated per number
of patients in active care at the 31th of December of each year. When reporting CD4
counts, unless otherwise indicated, only results that were obtained within 6 months of
the outcome of interest were included.
Ethics
The study protocol received exemption from ethical review from the Higher Degrees
Research and Ethics Committee from the School of Public Health of Makerere
University, because it involved only archived data that was stripped from identifiers.

53
The study received approval and was registered by the Uganda National Council of
Science and Technology (registration number SS3132).
Results
On 1 May 2001, ROM started with 10 HIV-infected patients. By 31 December 2012,
4784 patients were in active care, of which 92% were ≥18 years. In total, 1249
patients were classified ‘dead’ over the years, and of these, 1128 (90%) clinic files
were retrieved and entered into the database. None of the files from the 20 deaths
that occurred in 2001 were retrievable. Therefore, our study period covers 1 January
2002 to 31 December 2012.
After reviewing the database the study team excluded 100 deaths from further
analysis because they were <18 years of age (n=43) or their date of death was
unknown (n=57). Hence, we analyzed the data of 1028 deaths. At least one CD4 count
was available for 68% of patients that died after 2004. Over time, the mortality ratio
dropped from 15.4 deaths/100 adults in active care in 2002 to 2.3/100 adults in
active care in 2012.
Characteristics of the deaths
The median age of all deceased was 36 years (IQR 30-42), 58% were female, the
median duration between registration and death was 158 days (IQR 61-420), 35% of
all deaths occurred within 3 months after registration and 54% within 6 months
(Table 1). The median CD4 count prior to death was 90 cells/mm3 (IQR 22-237).
Thirty-eight percent (n=395) of deceased were on ART and 42% died within the first
3 months of ART, 56% in the first 6 months and 68% within the first year. When
observing CD4 count in the 6 months prior to death stratified by year of death, the
median CD4 count increased from 49 cells/mm3 in 2004 to 132 cells/mm3 in 2009
(Table 1). After 2010, it varied between 93 to 180 cells/mm3. The proportion of
deceased patients on ART steadily increased from 41% in 2004 to 61% in 2012. The
median duration between the start of ART and death was 61 days (IQR 37-119) in
2004, then increased to a maximum of 722 days (IQR 77-1501) in 2011 and decreased
again to 474 days (IQR 104-1118) in 2012.
Causes of death over time
Of the 1028 deceased, 784 (76%, 95%CI 74-79) died of a ‘communicable condition or
AIDS-defining malignancy’, 48 (5%, 95%CI 4-7) of a ‘chronic non-communicable
condition’, 29 (3%, 95%CI 2-4) of an ‘other non-communicable condition’, 47 (5%,
95%CI 3-6) of 2 ‘communicable conditions and/or AIDS-defining malignancies’ and 17
(2%, 95%CI 1-3) of a combination including at least one ‘communicable condition or
AIDS-defining malignancy’ and one ‘chronic non-communicable condition’. For 103
(11%, 95%CI 9-13) no cause of death could be determined.

54
Over time, the proportion of deaths from communicable conditions and AIDS-defining
malignancies showed a significant decrease, from 84% (95%CI 74-90) in 2002 to 64%
(95%CI 53-74) in 2012 (p<0.01). Death from chronic non-communicable conditions
showed a significant increase from 1% (95%CI 0-7) in 2002 to 7% (95%CI 3-16) in
2012 (p<0.01) as did death from other non-communicable conditions (from 1%
(95%CI 0-7) in 2002 to 7% (95%CI 3-16) in 2012, p<0.01) and death from a
combination of a communicable and non-communicable condition (from 0% in 2002
to 6% (95%CI 2-14) in 2012, p<0.01) (Table 2, Figure 1). The proportion of unknown
causes of death remained stable over time (p=0.16).
Overall, the main cause of death was TB, which was identified in 34% (95%CI 31-37)
of deceased (Table 2). There was a significant decrease in TB as cause of death
overtime from 43% (95%CI 32-53) in 2002 to a steady proportion of approximately
25% from 2006 onwards (p<0.01). 70% (95%CI 64-74) of the TB cases were on anti-
TB treatment. None of the other ‘communicable conditions and AIDS-defining
malignancies’ showed any clear trends in time.
Chronic medical conditions were identified in 6% (95%CI 5-8) of deaths overall. The
main chronic non-communicable cause of death was non AIDS-defining malignancy,
which showed a peak prevalence of 7% (95%CI 3-14%) in 2011. Esophageal and
breast cancer were the most frequent occurring non AIDS-related malignancies.
Accidents, suicide and assault were the main ‘other non-communicable condition’ and
caused death in 2% (95%CI 1-3).
Cause of death in patients on ART
In total, 395 patients died while on ART: 268 (68%) within 1 year of starting ART and
127 (32%) while on ART longer than 1 year. The median CD4 count of those dying
within 1 year of ART was 56 cells/mm3 (IQR 14-134) compared to 229 cells/mm3 (IQR
145-359) in those dying while on ART longer than 1 year (p<0.01).
Of the 268 patients that died within 1 year of ART initiation, 208 (78%, 95%CI 72-82)
died of a ‘communicable condition or AIDS-defining malignancy’, 8 (3%, 95%CI 2-6) of
a ‘chronic non-communicable condition’, 3 (1%, 95%CI 0-3) of an ‘other non-
communicable condition’, 29 (11%, 95%CI 8-15) of 2 ‘communicable conditions
and/or AIDS-defining malignancies’ and 3 (1%, 95%CI 0-3) of a combination of a
‘communicable condition or AIDS-defining malignancy’ and a ‘chronic non-
communicable condition’. In 17 deceased (6%, 95%CI 4-10) no cause of death was
identified. The main causes of death were TB (32%, 95%CI 26-38), unspecified
infection (19%, 95%CI 15-24), infection with respiratory symptoms (12%, 95%CI 9-
16) and Kaposi sarcoma (11%, 95%CI 8-15) (Table 3).
Of the 127 patients that died while on ART for > 1 year, 62 (49%, 95%CI 40-57) died
of a ‘communicable condition or AIDS-defining malignancy’, 18 (14%, 95%CI 9-21) of

Complete thesisOpmaakV7

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (15)

Similar to Complete thesisOpmaakV7

Similar to Complete thesisOpmaakV7 (20)

Complete thesisOpmaakV7