Skip to main content
PMC home page
PLOS ONE logoLink to PLOS ONE
. 2020 Feb 25;15(2):e0218695. doi: 10.1371/journal.pone.0218695

Severe anaemia complicating HIV in Malawi; Multiple co-existing aetiologies are associated with high mortality

Minke H W Huibers 1,2,*, Imelda Bates 3, Steve McKew 3,4, Theresa J Allain 5, Sarah E Coupland 6,7, Chimota Phiri 5, Kamija S Phiri 8, Michael Boele van Hensbroek 1, Job C Calis 1,9,10
Editor: Kostas Pantopoulos11
PMCID: PMC7041863  PMID: 32097440

Abstract

Background

Severe anaemia is a major cause of morbidity and mortality in HIV-infected adults living in resource-limited countries. Comprehensive data on the aetiology are lacking but are needed to improve outcomes.

Methods

HIV-infected adults with severe (haemoglobin ≤70g/l) or very severe anaemia (haemoglobin ≤ 50 g/l) were recruited at Queen Elizabeth Central Hospital, Blantyre, Malawi. Fifteen potential causes and associations with anaemia severity and mortality were explored.

Results

199 patients were enrolled: 42.2% had very severe anaemia and 45.7% were on ART. More than two potential causes for anaemia were present in 94% of the patients including iron deficiency (55.3%), underweight (BMI<20: 49.7%), TB infection (41.2%) and unsuppressed HIV infection (viral load >1000 copies/ml) (73.9%). EBV/CMV co-infection (16.5%) was associated with very severe anaemia (OR 2.8 95% CI 1.1–6.9). Overall mortality was high (53%; 100/199) with a median time to death of 17.5 days (IQR 6–55) days. Death was associated with folate deficiency (HR 2.2; 95% CI 1.2–3.8) and end stage renal disease (HR 3.2; 95% CI 1.6–6.2).

Conclusion

Mortality among severely anaemic HIV-infected adults is strikingly high. Clinicians should be aware of the urgent need for a multifactorial approach including starting or optimising HIV treatment, considering TB treatment, nutritional support and optimising renal management.

Introduction

Anaemia is recognized as the most common haematological complication of Human Immunodeficiency Virus (HIV) infection worldwide [1,2] The World Health Organization (WHO) defines anaemia as a haemoglobin level below 110–120 g/l. In sub-Saharan Africa, 60% of HIV-infected adults are anaemic and 22% are severely anaemic [3,4]. Anaemia is associated with an increased one-year mortality in HIV infection of 8%, which rises to 55% in those with severe anaemia [5,6]. Anaemia treatment may even improve survival [7]

To prevent and treat severe anaemia in HIV-infected patients, a comprehensive understanding of the aetiology and pathophysiology is essential. Severe anaemia in HIV infection has been associated with micronutrient deficiencies, infections (viral, bacterial and parasitic) and inflammation, medication induced (zidovudine and cotrimoxazole) and neoplastic diseases [8–12]. Only a few studies have comprehensively studied the multifactorial aetiology and pathophysiology of HIV-associated severe anaemia in sub-Saharan Africa despite the high burden of HIV infection in this region [2]. Commonly, studies only report on the association between HIV infection and a single cause of severe anaemia, for example iron deficiency, without considering the multiple causes of severe anaemia that may impact on an HIV-infected patient [13,14]. As a consequence, evidence to inform preventive or treatment guidelines for severe anaemia in HIV-infected patients in sub-Saharan Africa is scarce. In practice, severe anaemia management in HIV-infected patients is often still based on the same strategies used for non-HIV infected patients, including iron supplementation, malaria treatment and de-worming [2,15]. These strategies may be ineffective as the causes of HIV-associated severe anaemia may be different, and even harmful, since iron supplementation may exacerbate infections and potentially cause deterioration of the patient’s condition [16,17].

Better knowledge about the aetiology of severe anaemia is essential in order to develop evidence-based protocols and ultimately improve outcomes for HIV-infected adults with severe anaemia in sub-Saharan Africa. To address this knowledge gap, we performed a comprehensive observational cohort study to explore the prevalence of potential aetiologies of HIV-associated severe anaemia in Malawian in-patients, and studied associations between these and the severity of the anaemia and patient outcomes.

Methods

An observational cohort study of HIV-infected patients with severe anaemia (haemoglobin ≤70 g/l) admitted to the Queen Elizabeth Central Hospital (QECH), Blantyre, Malawi between February 2010 and March 2011 was performed. All HIV-infected patients above 18 years of age with severe anaemia admitted to the general medical ward were approached and enrolled in the study if they provided informed consent. A case record form including a detailed medical history and physical examination was completed for each enrolled patient. On admission a venous blood sample was collected and a chest X-ray was performed. The patients were managed according to the hospital protocols, which included a blood transfusion if required, treatment of correctable conditions including anti-malarial medication and antibiotics. Antiretroviral treatment (ART) was provided according to the national Malawi guidelines, which stipulated that ART should only be initiated by specialist outpatient clinicians, so patients were not started on ART on in-patient wards. At the time of the study first line ART was a combination of stavudine, lamivudine and nevirapine and second line treatment was a combination of zidovudine, lamivudine, tenofovir and lopinavir/ritonavir [18]. ART was only prescribed for WHO stage 3 and 4 disease, or WHO stage 1 and 2 disease with CD4 count < 350 x 109/l [18,19]. Cotrimoxazole was prescribed routinely to all HIV-infected patients on ART as Pneumocystis Jirovecii prophylaxis. For this study patients were followed up in a dedicated ART clinic after discharge. Follow-up was done for a maximum of 365 days after enrolment, or when they attended the ART clinic for routine appointments or, if they failed to attend, by a home visit from a study nurse.

Laboratory assays

All samples were analysed within 24 hours of collection or stored at -80°C for further analysis. On enrolment haemoglobin concentrations were measured on the ward using the HemoCue B-Haemoglobin analyser (HemoCue, Ängelholm, Sweden) to screen for eligibility. For patients enrolled in the study, the haemoglobin and red cell indices (MCV, MCH and MCHC) were determined using an automated analyser (Beckman Coulter, Durban, South Africa). CD4-cell counts were assessed using BD FACS Count (BD Biosciences, San Jose, CA, USA). Transferrin, iron, ferritin, folate and vitamin B12 were analysed on Modular P800 and Monular Analytic E170 systems (Roche Diagnostics, Switzerland). Soluble transferrin receptor (sTfR) levels were measured using ELISA (Ramco Laboratories, TX, USA). Serum creatinine was analysed using a Beckman Coulter CX5 (ADVIA 2400 Siemens Healthcare Diagnostics). Renal function was measured by estimating glomerular filtration rate (eGFR) using simplified Modification of Diet in Renal Disease (MDRD)-Study formula and the GFR was classified by Chronic Kidney Disease classification [20–22]. For all tests the manufacturers’ reference ranges were used; internationally accepted cut-offs were used to define deficiencies [23]. Thick blood films were prepared and stained for malaria microscopy. Malaria was defined as the presence of Plasmodium falciparum asexual parasites in the blood films. HIV infection was confirmed using two point-of-care antibody tests (Unigold® and Determine®). For blood cultures a venous blood sample was inoculated into BACTEC Myco/F-Lytic culture vials and incubated in a BACTEC 9050 automated culture system (Becton Dickinson) for 56 days. Sub-culturing blood and sputum, susceptibility testing and isolate identification were performed by standard techniques [23,24]. Likely contaminants were recorded as absence of pathogens. Sputum cultures were examined for mycobacteria using Ziehl–Nielsen staining. Whole-blood isolates were assessed for Epstein–Barr virus and cytomegalovirus infection by semi-quantitative PCR and for parvovirus B19 by real-time PCR [25]. All chest X-rays were reviewed by a radiologist for signs of pulmonary tuberculosis (TB). When TB was suspected, standardized treatment was started according to the local protocols.

Bone marrow

If the patient’s clinical condition allowed and they provided consent, a bone marrow aspirate and trephine biopsy were performed. All bone marrow samples were taken from the posterior iliac crest. Samples of the aspirates were spread onto slides and trephine biopsies were fixed, decalcified and embedded in paraffin wax [26, 27]. Bone marrow samples were sent to the Haematopathology Referral Centre at the Royal Liverpool University Hospital, Liverpool UK, for analysis. Sections of the trephine blocks were stained with haematoxylin and eosin, and Giemsa, with Perls stain for iron, and for reticulin [26]. Bone marrow analysis was performed by pathologists who were unaware of the patients’ data. All slides were examined for using a predefined format and diagnoses were allocated to the categories lymphoproliferative disease, myeloproliferative disease, myelodysplastic syndrome (MDS), TB and ‘other’ [28–30]. The need for additional histochemical (e.g. Ziehl-Neelsen) or immunohistological (e.g. CD3, CD20) staining was determined according to the local protocol in Liverpool depending on the preliminary morphological findings.

Definitions for potential factors involved in the aetiology of severe anaemia

A total of 15 potential factors involved in the aetiology of severe anaemia were investigated. Factors were based on two previous studies performed in Malawi and a systemic review on this topic [31–33].and comprised (with definitions) 1) Unsuppressed HIV-infection; viral load ≥1000 copies/ml. 2) TB: presence of one or more of the following: a) positive sputum culture, b) chest X-ray with signs of pulmonary TB and/or c) on going TB treatment at time of enrolment d) clinical diagnosis by local doctor including unknown generalized lymphadenopathy and/or night sweats of > 30 days and of unknown origin e) caseating granulomata in the bone marrow trephine. 3) Malaria: presence of malaria parasites in a thick blood film. 4) parvovirus B19: viral load of >1000 copies/ml. 5) Cytomegalovirus (CMV); viral load of >100 copies/ml. 6) Epstein-Barr virus (EBV); viral load of >100 copies/ml. 7) Bacteraemia; a blood culture growing a potential pathogen. 8) Underweight (BMI ≤18.5). 9) Serum folate deficiency (≤3 ng/l). 10) Vitamin B12 deficiency (≤180 pg/ml). 11). Iron deficiency; In a sub study (Huibers et al PLOS One 2019) we evaluated bone marrow (BM) iron deficiency using several conventional blood markers; MCV (fl), MCH (pg/cells), Fe (umol/l), ferritin (ug/dl), TFr1 receptor (nmol/l), TrF index (stFR/Log ferritin). All markers showed suboptimal correlations (i.e. AUCROC < 0.6) with BM iron deficiency, though MCV performed best (AUCROC 0.545). MCV was therefore used to identified iron deficiency as it was the best (though suboptimal) conventional marker in our setting. It was also available for patients beyond the subgroup (n = 76) who had a bone marrow result and it is available in most African settings [3, 22, 33]. Since an MCV <83fL is commonly used in other studies and guidelines [3, 22, 33] and it was the best predictor of BM iron deficiency in our setting, we used it as the marker for iron deficiency in this analysis. As the AUC-ROC was suboptimal we chose to call it MCV ≤83fL rather than iron deficiency. 12) Zidovudine usage. 13) Cotrimoxazole usage. 14) Bone marrow disorders; lymphoproliferative disease, myeloproliferative disease or MDS. 15) Renal impairment: a GFR which either indicated impaired (GFR 15–59 ml/min/1.73 m2) or End Stage (GFR ≤15 ml/min/1.73 m2) Renal Disease [22,35]].

Statistics

The study was primarily designed to give a complete overview of the potential factors associated with severe anaemia, including lymphomas, in severely anaemic African patients infected with HIV. The selected sample size of 200 would be able to detect an estimated prevalence of 5% or 10% with confidence intervals of 2.4% - 9% and 6–15% respectively. Baseline characteristics and prevalence of potential risk factors are presented as proportions or medians with IQR. Logistic regression was performed to model the association between anaemia and potential factors associated with anaemia. Results are expressed as OR with 95% CI and p-values. Variables associated with the outcome variables (P ≤ 0.10) in the univariate analysis were included in the multivariate model in a stepwise approach. Kaplan Meier survival curves were used to assess cumulative mortality. Significant differences were investigated with a Log Rank test. Uni- and multivariate analyses were done using logistic regression and Cox regression to describe predictors of overall mortality. A sub- analysis was performed for mortality within 60 days. A sub analysis was performed to evaluate the distribution of co-existing factors between mortality within and after 60 days. Group comparisons for categorical data were performed using the χ2 test or Fisher’s exact test, and for continuous data using the t-test or the Wilcoxon rank-sum test. P values of < 0.05 were regarded as statistically significant. All reported P values were two-sided. The data were analysed using Stata (version 12) (STATA Corp. LP, Texas, TX, USA).

Ethics

The Research Ethics Committee of the College of Medicine, University of Malawi (P.09.09.824) and the Research Ethics Committee of Liverpool School of Tropical Medicine (research protocol 09.64) approved the study. The purpose of the study was explained to the patients in the local language (Chichewa) and written informed consent was obtained before inclusion in the study.

Results

In total, 199 patients were included in the study: 64.8% were female. The median age was 32 years (IQR 27–61 years). The median haemoglobin was 53 g/l (IQR 4.2–6.3) and 84 (42%) patients had very severe anaemia (Hb ≤ 50 g/l). A total of 91 (45.7%) patients were on ART at enrolment including 79.1% on first line ART. During the study period, an additional 41 (21%) patients started on ART treatment. 67.1% of the patients were immune suppressed with a CD4 count ≤ 200 cells/mm. Baseline characteristics of the patients are shown in Table 1.

Table 1. Baseline characteristics of HIV-infected patients with severe anaemia at enrolment into study.

Overall
Age, years (median IQR) 32 (IQR 27–61)
Sex (female) 129/199 (64.8%)
Haematology
Haemoglobin (Hb) (median, IQR) g/l 53 (IQR 42–63)
Severe anaemia (Hb 51–70 g/l)
Very severe anaemia (Hb ≤ 50g/l)		 115/199 (57.8%)
84/199 (42.2%)
Pancytopenia1 42 /184 (21.2%)
Mortality
Overall mortality (365 days)
Early mortality (60 days)
Days until death (median, IQR)		 101/199 (50.8%)
81/101 (80.2%)
17.5 days (6–55)
HIV-disease and treatment
CD4 (median, IQR)
CD4 ≤200 cells/mm3 		175 (IQR 55–825)
104/155 (67.1%)
Viral load ≥1000 copies/ml 136/184 (73.9%)
ART at enrolment
First-line ART
Second-line ART
Non-specified ART		 91/199 (45.7%)
72/199 (36.2%)
11/199 (5.5%)
8/199 (4%)
Blood transfusions and supplemental treatment
Blood transfusion at enrolment 47/199 (23.6%)
Folate supplementation at enrolment 57/199 (28.6%)
Iron supplementation at enrolment 81/199 (40.7%)
Vitamin B12 supplementation at enrolment 0/199 (-)
Open in a new tab

1 Pancytopenia is defined as thrombocytopenia (≤150 x 109/l) and leucopenia (≤4 x 109/l) and severe anaemia (Hb ≤ 70 g/l) [22]. Abbreviations: Hb: Haemoglobin, ART: antiretroviral therapy.

The prevalence of factors and their association with severe anaemia is shown in Table 2. An unsuppressed HIV-infection occurred in 73.9% patients. TB was the second most common infection occurring in 82 (41.2%) of the patients. In 19 (23%) of these patients TB was diagnosed on their chest X-ray. Granulomata were seen in the bone marrow trephine in 15 (18%) of all the 82 patients that had a diagnosis of TB. 11 (13%) patients were on TB treatment at enrolment. 69/170 (40.5%) patients had evidence of current EBV infection and 57/170 (33.5%) had evidence of current CMV infection. Co-infection with CMV and EBV was found in 28/170 (16.5%) of the patients. Bacteraemia was diagnosed on a positive blood culture in 26 (13.1%); the most common pathogens were E. Coli (12 patients; 42.9%) and non-Typhoid Salmonella (5 patients; 19.9%). 74/148 (49.7%) patients were underweight and MCV ≤ 83 fl occurred in 61/180 (33.9%) of the patients. Bone marrow sampling was performed in 73 patients. Of these, 28 (38.4%) had morphological abnormalities with MDS being the most common abnormality (20 patients; 27.4%) (Table 2). Renal impairment was diagnosed in 36/185 patients (19.5%) and 12 of these patients (33%) had end stage renal disease. Overall, patients had a mean of 3 (range 1–8) co-existing aetiologies of severe anaemia (Fig 1). The overlap of factors is visually displayed in supplemental S1 Fig.

Table 2. Distribution and multivariate analysis of co-existing factors associated with severe (Hb≤ 70 g/l- Hb> 50 g/l) and very severe anaemia (Hb≤ 50 g/l) in HIV-infected adults in Malawi.

Overall N = 199(100%) Severe anaemia
N = 115/199 (57.8%)		 Very severe anaemia
N = 84/199 (42.2%)		 Univariate Multivariate
Odds 95% -CI P-value Odds 95% -CI P-value
Sex (female) 129 (64.8%) 74/115 (65.2%) 55/84 (65.4%) 1.1 0.6–1.9 0.869 - - -
HIV
CD4 ≤200 cells/mm3 104/155 (67.1%) 62/87 (71.3%) 42/68 (61.8%) 1.04 0.6–1.8 0.889 - - -
Viral load ≥1000 copies/ml 136/184 (73.9%) 81/104 (77.9%) 55/80 (68.8%) 0.6 0.3–1.2 0.163 - - -
On ART at enrolment 91/199 (42.7%) 52/115 (40.9%) 39/84 (45.2%) 1.1 0.6–1.8 0.865 - - -
Infection
Malaria 6/167 (1%) 3/100 (3.0%) 3/67 (4.5%) 1.5 0.3–7.8 0.617 - - -
Tuberculosis 82/199 (41.2%) 53/115 (46.0%) 29/84 (34.5%) 0.7 0.6–0.99 0.043 0.6 0.1–2.8 0.507
Bacteraemia1 26/199 (13.1%) 17/115 (61.7%) 9/82 (11.0%) 0.7 0.3–1.6 0.402 - - -
Parvovirus B19 7/170 (4.2%) 5/99 (5.0%) 2/71 (2.8%) 0.5 0.1–2.9 0.476 -
Cytomegalovirus (CMV) 57/170 (33.5%) 28/99 (28.3%) 29/71 (40.8%) 1.8 0.9–3.3 0.088 - -
Epstein-Barr virus (EBV) 69/170 (40.6%) 40/99 (40.4%) 29/71 (40.8%) 1.0 0.6–1.9 0.954 - - -
EBV/CMV co-infection 28/170 (16.5%) 12/99 (12.1%) 16/71 (22.5%) 2.1 0.9–4.5 0.075 2.8 1.2–7.0 0.024
Malnutrition
Underweight 74/148 (49.7%) 40/81 (49.3%) 34/68 (50.0%) 1.0 0.5–2.0 0.940 - - -
Vitamin B12 deficiency 2/194 (1.0%) 1/113 (0.8%) 1/81 (1.2%) 1.4 0.09–22.7 0.813 - - -
MCV ≤ 83 fl2 61/180 (33.9%) 38/104 (36.5%) 23/76 (30.3%) 0.8 0.40–1.42 0.380 - - -
Folate deficiency 23/194 (11.9%) 13/113 (11.5%) 10/84 (11.9%) 1.1 0.5–2.6 0.858 - - -
Medication
Cotrimoxazole 163/199 (81.9%) 99/115 (81.6%) 64/84 (76.2%) 0.5 0.3–2.3 0.076 0.5 0.2–1.2 0.120
Zidovudine 13/199 (6.5%) 5/50 (10.0%) 8/40 (20.0%) 2.3 0.7–7.5 0.187 - - -
Renal function
Impaired (GFR 15–60) 24/185 (13.0%) 12/105 (11.4%) 12/80 (15.0%) 1.5 0.6–3.6 0.339
End stage (GFR ≤15) 12/185 (6.5%) 3/105 (2.9%) 9/80 (11.3%) 4.6 1.2–17.6 0.027 4.0 0.9–16.8 0.061
Bone marrow
Bone marrow disease 28/71 (39.4%) 19/42 (45.2%) 9/29 (31.0%) 0.5 0.2–1.5 0.231 - -
Aetiology
Co-existing aetiologies per patient (mean, SD) 3.3 (1.3) 3.3 (1.2) 3.2 (1.4) 0.8 0.4–1.8 0.605 - - -
Open in a new tab

1 A total of 28-blood cultures were positive, the most common organisms were E. coli (42.9%; 12/28) and non-Typhoid Salmonella (17.9%; 5/28). 2 Iron deficiency was defined by MCV≤ 83 fl. Explanatory variables associated with the outcome variables (P > 0.10) in the univariable analysis were excluded in the multivariable model in a stepwise approach (-). Abbreviation: GFR; Glomerular filtration rate.

Fig 1. Total number of aetiologies for severe anaemia co-existing in each patient (n = 199).

Fig 1

Open in a new tab

Mean is 3 factors (SD 1.3), range 1–8. Aetiologies for severe anaemia include: 1) Unsuppressed HIV-infection; viral load ≥1000 copies/ml. 2) TB: one or more of the following were present: a) positive sputum culture, b) chest X-ray with signs of pulmonary TB and/or c) on going TB treatment at time of enrolment d) clinical diagnosis by local doctor including unknown generalized lymphadenopathy and/or night sweats of > 30 days and of unknown origin e) caseating granulomata in the bone marrow trephine. 3) Malaria: presence of malaria parasites in a thick blood film. 4) Parvovirus B19: viral load of >1000 copies/ml. 5) Cytomegalovirus (CMV); load of >100 copies/ml. 6) Epstein-Barr virus (EBV); viral load >100 copies/ml. 7) Bacteraemia; a blood culture growing a potential pathogen. 8) Underweight (BMI ≤18.5). 9) Serum folate deficiency (≤3 ng/l). 10) Vitamin B12 deficiency (≤180 pg/ml). 11). Iron deficiency defined by MCV ≤ 83 fl. 12) Zidovudine usage. 13) Cotrimoxazole usage. 14) Bone marrow disorders; lympho-proliferative disease, myeloid-proliferative disease or MDS. 15) Renal impairment: a GFR which either indicated impaired (GFR 15–59 ml/min/1.73 m2) or End Stage (GFR ≤15 ml/min/1.73 m2) Renal Disease [22, 35].

Comparing the different risk factors between very severe anaemia (Hb < 50g/L) and severe anaemia, EBV/CMV co-infection (OR 2.8 95% CI 1.1–6.9) was the only factor associated with very severe anaemia (Table 2).

During the one-year follow-up period, 101 study patients (50.8%) died. The median time to death was 17.5 days (IQR 6–55) and 81 (80.2%) of these deaths occurred within 60 days of admission (Fig 2). The factors associated with mortality within 60 days and those associated with overall mortality in HIV-infected adults in Malawi are provided in S1 Table. Folate deficiency and end stage renal disease were associated with overall mortality with Hazard Ratio 2.0 (95% CI 1.2–3.6) and Hazard Ratio 3.0 (95% CI 1.5–5.9) respectively (Fig 3). Only end stage renal disease (Hazard Ratio 2.7 95% CI 1.2–6.2) was associated with mortality within 60 days (S2 Table). Neither very severe anaemia (haemoglobin ≤50 g/l) nor the haemoglobin levels were associated with mortality in the study patients (Hazard Ratio 0.9, 95%CI 0.6–1.4 and Hazard Ratio 1.01, 95% CI 0.9–1.2 respectively).

Fig 2. Kaplan Meyer survival curve over time (days) for adult Malawian patients with HIV infection and severe anaemia during 365 days follow-up.

Fig 2

Open in a new tab

Abbreviation: 95% confidence interval (95% CI).

Fig 3. Risk factors for 365-day mortality in HIV-infected patients with severe anaemia.

Fig 3

Open in a new tab

Univariate and multivariate Cox regression outcome (Hazard Ratios 95% CI). Folate deficiency (≤3 ng/l) HR 2.0 95% CI 1.2–3.6 and end stage renal disease (GFR ≤15); HR 3.0 5% CI 1.5–5.9, were associated with overall mortality. Abbreviations: Hb: Haemoglobin, GFR; Glomerular filtration rate, VL Viral Load.

Discussion

In this study we described the prevalence of several potential aetiologies for severe and very severe anaemia in HIV-infected Malawian adults. Patients had a mean of three co-existing aetiologies potentially contributing to their anaemia, the most common one being unsuppressed HIV infection. Mortality in the study patients was extremely high, as 51% of the patients died within one year and most died within 60 days of admission. Severe anaemia in HIV-infected patients in a resource limited setting, such as Malawi, is therefore a multi-causal critical condition associated with high mortality.

Anaemia in HIV-infected patients is an independent predictor of mortality with deaths increasing as haemoglobin concentrations decrease [6, 35]. However HIV pathogenesis is complex and although anaemia may contribute to morbidity and mortality, the data suggest that the aetiology is multi-causal. Previous studies have reported an estimated one-year mortality of 30–55% in severely anaemic, compared to 3.7% in non-anaemic, HIV-infected patients in resource limited settings [5, 6]. Our study results are consistent with these outcomes. The seriousness of severe anaemia and its complexity in HIV-infected patients needs to be better recognized. Whether severe anaemia is the direct cause of the mortality or indirectly responsible through other contributing factors, severe anaemia should be recognized by clinicians as a sign of critical HIV disease and should therefore should be classified as a stage 4 condition instead of its current stage 3 classification [36, 37].

Unsuppressed HIV virus was present in 79% of the patients in our study population. HIV may cause anaemia directly through an inhibitory effect of the HIV-virus on the erythropoietin progenitor cells in the bone marrow, or indirectly through opportunistic infections and/or inflammation causing anaemia [38]. A large cross-sectional study in Tanzania showed that the risk of developing severe anaemia in HIV-infected patients was increased two- to three-fold among patients with advanced HIV disease [3]. Our findings, and those of previous studies, indicate that controlling HIV infection by starting or switching ART treatment should be prioritised as the most important and urgent step in treatment protocols for severely anaemic HIV-infected patients. After our study had been completed guidelines for initiating ART changed. At the time of the study the trigger for starting ART was based on a patient’s CD4 count, whereas it is currently recommended that ART should be started early in the course of HIV disease [18, 37]. It will be important for the impact of this policy change on HIV-related anaemia, and its consequences, to be evaluated. Irrespective of the policy change for initiating ART, the findings of our study remain very relevant because many HIV-infected patients in resource-limited settings present late in the course of their disease or are unable to access reliable supplies of ART. Consequently this patient population is likely to experience high levels of life-threatening anaemia since it is closely associated with severe HIV disease and may be exacerbated by the effect of medication such as zidovidine and cotrimoxazole on the bone marrow [39, 40].

TB has previously been associated with anaemia in HIV-infected patients [40]. In our study patients TB was a common co-infection with HIV (41%). This prevalence is comparable to previous reports on TB prevalence (43%) among severely anaemic HIV-infected patients in Africa [9, 15]. The pathophysiology of TB-associated anaemia in HIV-infected patients remains unclear. Bone marrow invasion by TB organisms or altered iron metabolism, as a side effect of tuberculostatic drugs, have been described [41]. Only 13% of patients were on TB treatment at enrolment. TB medication itself was not associated with the severity of anaemia (OR 1.6 95%CI 0.7–3.8). Given that we and others have found TB in nearly half of the HIV-infected patients with severe anaemia, TB screening and rapid initiation of treatment should be a high priority in the management of these patients, especially in resource limiting settings where there is a high TB prevalence.

Viral infections such as EBV, CMV and parvovirus B19 have been associated with anaemia in HIV-infected patients [42, 43]. In our study EBV and CMV were common and present in 40% and 35% of patients respectively, whilst parvovirus B19 was less common (4%). Although parvovirus B19 is pathophysiologically linked to mild anaemia, its role in the development severe anaemia in HIV infected patients has not been clearly established [43, 44]. Data on co-infections are very limited in African patients [33, 45, 46]. Our study is the first to describe the association between co-existing CMV and EBV infections and very severe anaemia (haemoglobin ≤50 g/l] which may be related to direct viral inhibition of erythropoiesis [45, 47, 48]. The majority of our patients with co-infection had advanced HIV disease (88.5%) but the association between EBV and CMV co-infection and very severe anaemia remained significant even after correction for advanced HIV disease. As CMV is a treatable infection it will be important to determine the effect of ART on CMV infection and severe anaemia, as ART can reduce CMV infection by improving immune status [49].

Malaria is unsurprisingly associated with anaemia in HIV-infected patients in sub-Saharan Africa [2, 15]. However, the contribution of malaria to anaemia in our study population was small as only 6 patients (3.5%) had malaria parasites on enrolment. This is in line with data from other studies that show that the role of malaria in causing anaemia, especially in HIV patients in sub-Saharan Africa [2, 14], is limited and likely to have been overestimated [31, 33].

Malnutrition was common in our population as half the patients had a BMI below 18.5m2 and deficiencies of iron and folate were diagnosed in 33.9% and 12% of the patients respectively. The prevalence of iron deficiency (MCV ≤ 83 fl) in our study is higher than the 18%-25% prevalence reported in anaemic HIV-infected patients in sub-Saharan Africa reported in previous studies [50, 51]. Interestingly, children and adults in sub-Saharan Africa with severe anaemia but without HIV infection, have a higher prevalence of iron deficiency of respectively 59% and 47%, than we found in our HIV-infected population [30,33]. Previous studies on severe anaemic HIV patients were published before 2010; at that time ART availability was limited and all the studies included HIV-infected patients who were ART naïve and who had high rates of immune suppression. HIV-infected patients who are in the early stages of infection or who are on effective treatment may have better immune function so in these patients the aetiology of severe anaemia (e.g. iron deficiency) may be more similar to that of non-HIV infected patients [33]. Iron deficiency diagnosis is complicated by a lack of peripheral iron markers and microcytosis (i.e. low MCV) poorly reflects bone-marrow iron deficiency especially in the context of HIV-infection (Huibers et al PLOS One, 2019). Bone marrow evaluation remains the golden standard for diagnosis of iron deficiency. In addition to finding a high prevalence of iron deficiency (MCV g 83 fl), ours is the first study to document a possible association between folate deficiency and increased mortality in severely anaemic HIV-infected patients. There are a limited number of studies describing micronutrient supplementation, including folate, for anaemic HIV-infected adults, but overall supplementation appears to have little effect on reducing morbidity and mortality [43,52]. Macronutrient support using, for example, fortified wheat flour, has had beneficial effects on anaemia reduction and micronutrient levels in populations in sub-Saharan Africa, but has never been tested in the context of severe anaemia in HIV-infected patients [53,54]. More research is therefore needed to evaluate the effectiveness of both macro- and micro- nutritional support in severely anaemic HIV-infected adults.

HIV-infected adults have an increased risk of neoplastic bone marrow diseases, which can often cause anaemia [14]. However, prevalence data for such conditions among HIV-infected patients with anaemia, especially in resource limiting settings, are scarce. Only three of our study patients had confirmed bone marrow malignancies. In contrast, and in line with previous studies, MDS was common occurring in 27% of our patients [28,55].

Renal impairment was a frequent finding among our study patients and has been linked to HIV disease progression, anaemia and poor outcomes in both wealthy and resource limited settings [56–59]. Evaluation of renal function is an important component of severe anaemia treatment protocols [34, 57] since it may affect the choice of ART [56, 60]. It may also have implications for clinical management, for example by introducing measures to prevent further deterioration or considering the use of erythropoietin[57].

Our study has several limitations. We purposely did not explore factors potentially associated with anaemia that previous studies had shown were uncommon in Malawi, such as haemoglobinopathies and parasitic infections [33, 61]. We also did not include an HIV-infected population without severe anaemia against which to compare the prevalence of anaemia aetiologies and clinical outcomes. Only a sub-population of our patients gave consent for bone marrow sampling. Although these were an unselected group of patients it is possible that this may have introduced a bias, for example by excluding the patients who were particularly unwell. Nevertheless, our findings are very valuable since bone marrow data from HIV-infected African patients is very scarce.

Conclusion

Our study has demonstrated that severe anaemia in HIV-infected adults in Malawi is associated with multiple co-existing aetiologies and has a strikingly high mortality rate. Severe anaemia in HIV-infected patients is therefore a critical indicator of mortality and requires urgent and multiple interventions. Particularly important are the initiation of ART, the management of infections such as TB and CMV, and optimisation of renal function. Intervention studies are needed to properly define the role and safety of iron and folate supplementation, as well as to develop and evaluate guidelines, which are feasible in resource-limited settings to help clinicians manage these patients more effectively.

Supporting information

S1 Data

(DTA)

Click here for additional data file. (92.3KB, dta)
S1 Fig

(TIF)

Click here for additional data file. (1.8MB, tif)
S1 Table

(DOCX)

Click here for additional data file. (94KB, docx)
S2 Table

(DOCX)

Click here for additional data file. (85.9KB, docx)

Acknowledgments

The authors would like to thank all of the study participants, doctors, nurses and support staff of Queens Elizabeth Hospital and the Malawi-Liverpool-Wellcome centre in Blantyre for their participation and cooperation. This study was supported by the Nutricia Research Foundation (Project number 2017–43), The Hague, the Netherlands and the Wellcome Trust (Project number WT086559), Liverpool, United Kingdom. The funders had no role in the study design, data collection and analysis, decision to publish or preparation of the manuscript.

Data Availability

All relevant data are within the manuscript and its Supporting Information files.

Funding Statement

Dr. Steve McKew received a grant from the Wellcome Trust fund; Project number: WT086559, Liverpool, United Kingdom grant. Data collection and laboratory testing were all paid by this grant. Drs. M. Huibers received a grant from the Nutricia research foundation; Project number 2017-43 The Hague, The Netherlands. All other authors were funded by there own institutions. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

  • 1.Mocroft A, Kirk O, Barton SE, Dietrich M, Proenca R, Colebunders R, et al. Anaemia is an independent predictive marker for clinical prognosis in HIV-infected patients from across Europe. EuroSIDA study group. AIDS (London, England). 1999;13(8):943–50. [DOI] [PubMed] [Google Scholar]
  • 2.Marti-Carvajal AJ, Sola I, Pena-Marti GE, Comunian-Carrasco G. Treatment for anemia in people with AIDS. Cochrane Database Syst Rev. 2011(10):Cd004776 10.1002/14651858.CD004776.pub3 [DOI] [PubMed] [Google Scholar]
  • 3.Makubi A, Okuma J, Spiegelman D, Hawkins C, Darling AM, Jackson E, et al. Burden and Determinants of Severe Anemia among HIV-Infected Adults: Results from a Large Urban HIV Program in Tanzania, East Africa. J Int Assoc Provid AIDS Care. 2015;14(2):148–55. 10.1177/2325957413488195 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Petraro P, Duggan C, Spiegelman D, Hertzmark E, Makubi A, Chalamilla G, et al. Determinants of Anemia Among Human Immunodeficiency Virus-Positive Adults at Care and Treatment Clinics in Dar es Salaam, Tanzania. The American journal of tropical medicine and hygiene. 2016;94(2):384–92. 10.4269/ajtmh.15-0587 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Santiago-Rodriguez EJ, Mayor AM, Fernandez-Santos DM, Ruiz-Candelaria Y, Hunter-Mellado RF. Anemia in a cohort of HIV-infected Hispanics: prevalence, associated factors and impact on one-year mortality. BMC Res Notes. 2014;7:439 10.1186/1756-0500-7-439 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Johannessen A, Naman E, Ngowi BJ, Sandvik L, Matee MI, Aglen HE, et al. Predictors of mortality in HIV-infected patients starting antiretroviral therapy in a rural hospital in Tanzania. BMC infectious diseases. 2008;8:52 10.1186/1471-2334-8-52 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Sullivan P. Associations of anemia, treatments for anemia, and survival in patients with human immunodeficiency virus infection. J Infect Dis. 2002. May 15;185 Suppl 2:S138–42. [DOI] [PubMed] [Google Scholar]
  • 8.Shu T, Jing C, Lv Z, Xie Y, Xu J, Wu J. Hepcidin in tumor-related iron deficiency anemia and tumor-related anemia of chronic disease: pathogenic mechanisms and diagnosis. Eur J Haematol. 2015;94(1):67–73. 10.1111/ejh.12402 [DOI] [PubMed] [Google Scholar]
  • 9.Butensky E, Kennedy CM, Lee MM, Harmatz P, Miaskowski C. Potential mechanisms for altered iron metabolism in human immunodeficiency virus disease. J Assoc Nurses AIDS Care. 2004;15(6):31–45. 10.1177/1055329004271125 [DOI] [PubMed] [Google Scholar]
  • 10.Kupka R, Msamanga GI, Mugusi F, Petraro P, Hunter DJ, Fawzi WW. Iron status is an important cause of anemia in HIV-infected Tanzanian women but is not related to accelerated HIV disease progression. J Nutr. 2007;137(10):2317–23. 10.1093/jn/137.10.2317 [DOI] [PubMed] [Google Scholar]
  • 11.Skikne BS, Punnonen K, Caldron PH, Bennett MT, Rehu M, Gasior GH, et al. Improved differential diagnosis of anemia of chronic disease and iron deficiency anemia: a prospective multicenter evaluation of soluble transferrin receptor and the sTfR/log ferritin index. American journal of hematology. 2011;86(11):923–7. 10.1002/ajh.22108 [DOI] [PubMed] [Google Scholar]
  • 12.Klassen MK, Lewin-Smith M, Frankel SS, Nelson AM. Pathology of human immunodeficiency virus infection: noninfectious conditions. Ann Diagn Pathol. 1997;1(1):57–64. 10.1016/s1092-9134(97)80009-3 [DOI] [PubMed] [Google Scholar]
  • 13.Akilimali PZ, Kashala-Abotnes E, Musumari PM, Kayembe PK, Tylleskar T, Mapatano MA. Predictors of Persistent Anaemia in the First Year of Antiretroviral Therapy: A Retrospective Cohort Study from Goma, the Democratic Republic of Congo. PloS one. 2015;10(10):e0140240 10.1371/journal.pone.0140240 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Levine AM. Anemia in the setting of cancer and human immunodeficiency virus. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2003;37 Suppl 4:S304–14. [DOI] [PubMed] [Google Scholar]
  • 15.Brentlinger PE, Silva WP, Vermund SH, Valverde E, Buene M, Moon TD. Practical Management of HIV-Associated Anemia in Resource-Limited Settings: Prospective Observational Evaluation of a New Mozambican Guideline. AIDS research and human retroviruses. 2016;32(1):12–25. 10.1089/AID.2015.0030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Sazawal S, Black RE, Ramsan M, Chwaya HM, Stoltzfus RJ, Dutta A, et al. Effects of routine prophylactic supplementation with iron and folic acid on admission to hospital and mortality in preschool children in a high malaria transmission setting: community-based, randomised, placebo-controlled trial. Lancet (London, England). 2006;367(9505):133–43. [DOI] [PubMed] [Google Scholar]
  • 17.Esan MO, van Hensbroek MB, Nkhoma E, Musicha C, White SA, Ter Kuile FO, et al. Iron supplementation in HIV-infected Malawian children with anemia: a double-blind, randomized, controlled trial. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2013;57(11):1626–34. [DOI] [PubMed] [Google Scholar]
  • 18.Ministry of Health Malawi. Treatment of AIDS, guidelines for the use of antiretroviral therapy in Malawi, edition 2008.
  • 19.World Health Organisation (WHO). Antiretroviral therapy for HIV infection in adults and adolescents; recommendations for a public health approach. 2006 revision2006 [Available from: https://www.who.int/hiv/pub/guidelines/artadultguidelines.pdf?ua=1. [PubMed]
  • 20.Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, et al. A new equation to estimate glomerular filtration rate. Annals of internal medicine. 2009;150(9):604–12. 10.7326/0003-4819-150-9-200905050-00006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Levey AS, Coresh J, Greene T, Stevens LA, Zhang YL, Hendriksen S, et al. Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Annals of internal medicine. 2006;145(4):247–54. 10.7326/0003-4819-145-4-200608150-00004 [DOI] [PubMed] [Google Scholar]
  • 22.Foundation NK. K/DOQI clinical practice guidelines for chronic kidney disease (2002). Retrieved 2008-06-29. 2008. [Google Scholar]
  • 23.Bates I. Practical hematology. Chapter 2: reference ranges and normal values. 2018. [Google Scholar]
  • 24.Michael L. Wislon MM, Morris Author J., Muray Patrick R., Reimer Larry G., Reller L. Barth, Towns Michael, Weinstein Melvin P., Wellstood Sybil A., Dunne W. Michael, Jerris Robert C., Welch David F. Principles and procedures of blood cultures; approved guideline (M47-A). Clinical and laboratory standards institute. 2007;27(17). [Google Scholar]
  • 25.Phiri KS, Calis JC, Kachala D, Borgstein E, Waluza J, Bates I, et al. Improved method for assessing iron stores in the bone marrow. Journal of clinical pathology. 2009;62(8):685–9. 10.1136/jcp.2009.064451 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Bain BJ. Bone marrow trephine biopsy. Journal of clinical pathology. 2001;54(10):737–42. 10.1136/jcp.54.10.737 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Bain BJ. Bone marrow aspiration. Journal of clinical pathology. 2001;54(9):657–63. 10.1136/jcp.54.9.657 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Garcia-Manero G. Myelodysplastic syndromes: 2014 update on diagnosis, risk-stratification, and management. American journal of hematology. 2014;89(1):97–108. 10.1002/ajh.23642 [DOI] [PubMed] [Google Scholar]
  • 29.Gale E, Torrance J, Bothwell T. The quantitative estimation of total iron stores in human bone marrow. J Clin Invest. 1963;42:1076–82. 10.1172/JCI104793 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Finch CA. The detection of iron overload. N Engl J Med. 1982;307(27):1702–4. 10.1056/NEJM198212303072709 [DOI] [PubMed] [Google Scholar]
  • 31.Lewis DK, Whitty CJ, Walsh AL, Epino H, Broek NR, Letsky EA, et al. Treatable factors associated with severe anaemia in adults admitted to medical wards in Blantyre, Malawi, an area of high HIV seroprevalence. Transactions of the Royal Society of Tropical Medicine and Hygiene. 2005;99(8):561–7. 10.1016/j.trstmh.2005.01.002 [DOI] [PubMed] [Google Scholar]
  • 32.Calis JC, van Hensbroek MB, de Haan RJ, Moons P, Brabin BJ, Bates I. HIV-associated anemia in children: a systematic review from a global perspective. AIDS (London, England). 2008;22(10):1099–112. [DOI] [PubMed] [Google Scholar]
  • 33.Calis JC, Phiri KS, Faragher EB, Brabin BJ, Bates I, Cuevas LE, et al. Severe anemia in Malawian children. N Engl J Med. 2008;358(9):888–99. 10.1056/NEJMoa072727 [DOI] [PubMed] [Google Scholar]
  • 34.Volberding PA, Levine AM, Dieterich D, Mildvan D, Mitsuyasu R, Saag M. Anemia in HIV infection: clinical impact and evidence-based management strategies. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2004;38(10):1454–63. [DOI] [PubMed] [Google Scholar]
  • 35.Gregory M. Lucas MJR, Stock Peter G., Shlipak Michael G., Wyatt Christina M., Gupta Samir K., Atta Mohamed G., Wools-Kaloustian Kara K., Pham Paul A., Bruggeman Leslie A., Lennox Jeffrey L., Ray Patricio E., Kalayjian Robert C. Clinical Practice Guideline for the Management of Chronic Kidney Disease in Patients Infected With HIV: 2014 Update by the HIV Medicine Association of the Infectious Diseases Society of America 2014. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2014;59(9):e96–e138. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.O'Brien ME, Kupka R, Msamanga GI, Saathoff E, Hunter DJ, Fawzi WW. Anemia is an independent predictor of mortality and immunologic progression of disease among women with HIV in Tanzania. Journal of acquired immune deficiency syndromes (1999). 2005;40(2):219–25. [DOI] [PubMed] [Google Scholar]
  • 37.World Health Organization (WHO). Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection. Recommendations for a public health aproach. Second edition. 2016 [Available from: https://www.who.int/hiv/pub/arv/arv-2016/en/. [PubMed]
  • 38.Wang Z, Goldberg MA, Scadden DT. HIV-1 suppresses erythropoietin production in vitro. Experimental hematology. 1993;21(5):683–8. [PubMed] [Google Scholar]
  • 39.Rivas P, Gorgolas M, Fernandez-Guerrero ML. Zidovudine and red-cell distribution width. N Engl J Med. 2005;352(20):2141–2. 10.1056/NEJM200505193522023 [DOI] [PubMed] [Google Scholar]
  • 40.Saadani Hassani A, Marston BJ, Kaplan JE. Assessment of the impact of cotrimoxazole prophylaxis on key outcomes among HIV-infected adults in low- and middle-income countries: a systematic review. Journal of acquired immune deficiency syndromes (1999). 2015;68 Suppl 3:S257–69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Kerkhoff AD, Meintjes G, Opie J, Vogt M, Jhilmeet N, Wood R, et al. Anaemia in patients with HIV-associated TB: relative contributions of anaemia of chronic disease and iron deficiency. Int J Tuberc Lung Dis. 2016;20(2):193–201. 10.5588/ijtld.15.0558 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Gedefaw L, Yemane T, Sahlemariam Z, Yilma D. Anemia and risk factors in HAART naive and HAART experienced HIV positive persons in south west Ethiopia: a comparative study. PloS one. 2013;8(8):e72202 10.1371/journal.pone.0072202 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Siegfried N, Irlam JH, Visser ME, Rollins NN. Micronutrient supplementation in pregnant women with HIV infection. Cochrane Database Syst Rev. 2012(3):Cd009755 10.1002/14651858.CD009755 [DOI] [PubMed] [Google Scholar]
  • 44.Morelli P, Bestetti G, Longhi E, Parravicini C, Corbellino M, Meroni L. Persistent parvovirus B19-induced anemia in an HIV-infected patient under HAART. Case report and review of literature. Eur J Clin Microbiol Infect Dis. 2007;26(11):833–7. 10.1007/s10096-007-0360-y [DOI] [PubMed] [Google Scholar]
  • 45.Bates M, Brantsaeter AB. Human cytomegalovirus (CMV) in Africa: a neglected but important pathogen. J Virus Erad. 2016;2(3):136–42. [PMC free article] [PubMed] [Google Scholar]
  • 46.Kelly MJ, Benjamin LA, Cartwright K, Ajdukiewicz KM, Cohen DB, Menyere M, et al. Epstein-barr virus coinfection in cerebrospinal fluid is associated with increased mortality in Malawian adults with bacterial meningitis. The Journal of infectious diseases. 2012;205(1):106–10. 10.1093/infdis/jir707 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Gumbo H, Chasekwa B, Church JA, Ntozini R, Mutasa K, Humphrey JH, et al. Congenital and postnatal CMV and EBV acquisition in HIV-infected Zimbabwean infants. PloS one. 2014;9(12):e114870 10.1371/journal.pone.0114870 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Olson D, Huntington MK. Co-infection with cytomegalovirus and Epstein-Barr virus in mononucleosis: case report and review of literature. S D Med. 2009;62(9):349, 51–3. [PubMed] [Google Scholar]
  • 49.Mofenson LM, Brady MT, Danner SP, Dominguez KL, Hazra R, Handelsman E, et al. Guidelines for the Prevention and Treatment of Opportunistic Infections among HIV-exposed and HIV-infected children: recommendations from CDC, the National Institutes of Health, the HIV Medicine Association of the Infectious Diseases Society of America, the Pediatric Infectious Diseases Society, and the American Academy of Pediatrics. MMWR Recomm Rep. 2009;58(Rr-11):1–166. [PMC free article] [PubMed] [Google Scholar]
  • 50.Lewis DK, Whitty CJ, Epino H, Letsky EA, Mukiibi JM, van den Broek NR. Interpreting tests for iron deficiency among adults in a high HIV prevalence African setting: routine tests may lead to misdiagnosis. Transactions of the Royal Society of Tropical Medicine and Hygiene. 2007;101(6):613–7. 10.1016/j.trstmh.2006.11.004 [DOI] [PubMed] [Google Scholar]
  • 51.Kagu MB, Khalil MI, Ahmed SG. Bone marrow macrophage iron stores in patients with HIV infection and AIDS-associated Kaposi's sarcoma. African journal of medicine and medical sciences. 2007;36(2):125–8. [PubMed] [Google Scholar]
  • 52.Semba RD, Ricketts EP, Mehta S, Netski D, Thomas D, Kirk G, et al. Effect of micronutrients and iron supplementation on hemoglobin, iron status, and plasma hepatitis C and HIV RNA levels in female injection drug users: a controlled clinical trial. Journal of acquired immune deficiency syndromes (1999). 2007;45(3):298–303. [DOI] [PubMed] [Google Scholar]
  • 53.Engle-Stone R, Nankap M, Ndjebayi AO, Allen LH, Shahab-Ferdows S, Hampel D, et al. Iron, Zinc, Folate, and Vitamin B-12 Status Increased among Women and Children in Yaounde and Douala, Cameroon, 1 Year after Introducing Fortified Wheat Flour. J Nutr. 2017;147(7):1426–36. 10.3945/jn.116.245076 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Noor RA, Abioye AI, Ulenga N, Msham S, Kaishozi G, Gunaratna NS, et al. Large -scale wheat flour folic acid fortification program increases plasma folate levels among women of reproductive age in urban Tanzania. PloS one. 2017;12(8):e0182099 10.1371/journal.pone.0182099 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Williamson BT, Leitch HA. Higher Risk Myelodysplastic Syndromes in Patients with Well-Controlled HIV Infection: Clinical Features, Treatment, and Outcome. Case Rep Hematol. 2016;2016:8502641 10.1155/2016/8502641 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Mulenga LB, Kruse G, Lakhi S, Cantrell RA, Reid SE, Zulu I, et al. Baseline renal insufficiency and risk of death among HIV-infected adults on antiretroviral therapy in Lusaka, Zambia. AIDS (London, England). 2008;22(14):1821–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Assaram S, Magula NP, Mewa Kinoo S, Mashamba-Thompson TP. Renal manifestations of HIV during the antiretroviral era in South Africa: a systematic scoping review. Syst Rev. 2017;6(1):200 10.1186/s13643-017-0605-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Husain NE, Ahmed MH, Almobarak AO, Noor SK, Elmadhoun WM, Awadalla H, et al. HIV-Associated Nephropathy in Africa: Pathology, Clinical Presentation and Strategy for Prevention. J Clin Med Res. 2018;10(1):1–8. 10.14740/jocmr3235w [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Fiseha T, Tamir Z, Seid A, Demsiss W. Prevalence of anemia in renal insufficiency among HIV infected patients initiating ART at a hospital in Northeast Ethiopia. BMC hematology. 2017;17:1 10.1186/s12878-017-0071-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Ministry of Health Malawi. 3rd Edition of the Malawi Guidelines for Clinical Management of HIV in Children and Adults 2016.
  • 61.McGann PT, Williams AM, Ellis G, McElhinney KE, Romano L, Woodall J, et al. Prevalence of inherited blood disorders and associations with malaria and anemia in Malawian children. Blood advances. 2018;2(21):3035–44. 10.1182/bloodadvances.2018023069 [DOI] [PMC free article] [PubMed] [Google Scholar]

Decision Letter 0

Kostas Pantopoulos

17 Jul 2019

PONE-D-19-15827

Severe anaemia complicating HIV in Malawi; multiple co-existing aetiologies are associated with high mortality.

PLOS ONE

Dear mrs Huibers,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Among them, particular attention should be given on comparing patients who died early, later or survived, and also on comparing patients on ART or not.

We would appreciate receiving your revised manuscript by Aug 31 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Kostas Pantopoulos, PhD

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

1. Please note that all PLOS journals ask authors to adhere to our policies for sharing of data and materials: https://journals.plos.org/plosone/s/data-availability. According to PLOS ONE’s Data Availability policy, we require that the minimal dataset underlying results reported in the submission must be made immediately and freely available at the time of publication. As such, please remove any instances of 'unpublished data' or 'data not shown' in your manuscript and replace these with either the relevant data (in the form of additional figures, tables or descriptive text, as appropriate), a citation to where the data can be found, or remove altogether any statements supported by data not presented in the manuscript.

2. We note that you have included the phrase “data not shown” in your manuscript. Unfortunately, this does not meet our data sharing requirements. PLOS does not permit references to inaccessible data. We require that authors provide all relevant data within the paper, Supporting Information files, or in an acceptable, public repository. Please add a citation to support this phrase or upload the data that corresponds with these findings to a stable repository (such as Figshare or Dryad) and provide and URLs, DOIs, or accession numbers that may be used to access these data. Or, if the data are not a core part of the research being presented in your study, we ask that you remove the phrase that refers to these data.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: General comments:

It is well established that anaemia and anaemia severity are associated with HIV disease stage and mortality. HIV-associated anaemia has many potential causes. In this manuscript, Huibers et al report the prevalence of diverse co-existing morbidities that can be linked to anaemia in a cohort of hospitalised HIV patients presenting with severe anaemia (Hb<7g/dL, N=199 – median 5.3g/dL), many of whom were immunosuppressed. They classified participants as having severe anaemia or very severe anaemia (Hb<5g/dL) and tested, using univariate and multivariate analyses, whether there were differences in risk of co-morbidities according to anaemia severity; they found some evidence of increased risk of CMV-EBV infection in the very severe anaemia group. The cohort had a high mortality rate, presumably reflecting a high prevalence of advanced HIV disease at hospital presentation. Mortality was associated with end-stage renal disease and (more weakly) with folate deficiency; anaemia severity was not associated with mortality risk (bearing in mind all participants had Hb<7g/dL according to enrolment criteria).

Similar studies from the same location have previously been published (e.g. Lewis et al, Trans R Soc Trop Med Hyg. 2005) considering multiple-potential aetiologies of anaemia in hospitalised patients. The data in the present study are largely descriptive, focussing purely on HIV-infected patients; they are robust, although the inclusion criterion of Hb<7g/dL does limit evaluation to comparison of severe vs very severe anaemia (as opposed to e.g. no anaemia/mild anaemia). The discussion is long and could be made more concise. In places it seems to infer a causative relationship between anaemia and mortality, but the present analyses do not allow such inferences to be made; HIV pathogenesis is complex, and although anaemia may contribute to morbidity and mortality, the data show that there are many other co-morbidities that could also explain the mortality risk.

This manuscript was submitted with a parallel manuscript evaluating markers of iron status in a subset of the above cohort of severely anaemic HIV-infected patients, to be evaluated concurrently.

Specific points:

1. Data availability: the authors state that raw data will be available within the manuscript or supplementary information – it’s not clear whether this has been provided yet. It would be helpful to refer to this within the manuscript.

2. Introduction (paragraph 2): it would be useful to include brief consideration of how important anaemia is clinically in relation to other co-morbidities (i.e. how important is correction of anaemia per se).

3. Introduction: The authors should specifically mention inflammation as a potential cause of anaemia (this will likely be present during many of the infections, but may not be restricted to these).

4. Methods: the authors confirm whether blinding was used in bone marrow analyses.

5. Methods/Results/Discussion (line 83): iron deficiency – the parallel manuscript demonstrates that MCV is a poor index of iron status (bone marrow iron staining, considered gold standard). The authors should make this clear in the present study in the text / table footnotes (e.g. by referring to the second manuscript) – it is therefore unclear how useful this part of the analysis is. Could the authors incorporate bone marrow iron status into analysis, for the subset of patients for whom that information is available?

6. Results – Table 2:

o % with severe anaemia (in title row) should be corrected.

o In the table footer, the authors should clarify that analyses are comparing severe with very severe anaemia.

o The methods section states that variables with P<0.1 in univariate are included in multivariate analysis: it appears that variables with P>0.1 are included – this should be clarified / corrected.

o Renal function: normal – is it intentional that no statistics are included here?

7. Results:

o do particular co-morbidities frequently occur together?

o do any particular combinations associate with anaemia severity, haemoglobin concentration or mortality/mortality within first 60 days?

8. Results: Table 1 / Text: there is a discrepancy between the mortality reported in the text and in the Table. This should be corrected.

9. Results: Given the pattern of mortality (approx. 50% within 1 year, approx. 80% of whom died within 60 days):

o The analysis could be refined to compare patients who died within 60 days vs those who died later and/or those who survived: was there a difference in anaemia group, haemoglobin concentration, or co-morbidities between these groups?

10. Results: Figure 3 – the position of the “zero” line should be aligned with 0.

11. Discussion:

o In general the discussion can be made more concise, e.g. see repetition in lines 10-11 and 20-21.

o Lines 14-15: this could be interpreted as anaemia being causal for mortality; it is part of a complex disease state. Figure 3 found no association of anaemia severity with mortality risk (noting that all individuals had at least severe anaemia). However, in concert with much previous literature, the data are consistent with a associations between severe anaemia and mortality in HIV infection given this mortality rate. Suggest revising to “a multi-causal critical condition associated with high mortality”.

o Line 26-27: it is a little unclear whether the authors are implying that severe anaemia is itself a prime cause of mortality (that therefore should be addressed per se), or rather whether it is a clear marker of severe disease (line 30) that would likely be improved through addressing co-existing conditions associated with advanced HIV (infections etc). The literature cited regarding anaemia and mortality is correlative, so the latter interpretation is favoured in my view.

o Line 34: add “and inflammation”?

o Line 36: the lack of effect here could reflect all patients having Hb<7g/dL in the present study.

o Line 39-41: the authors could cite relevant data on the impact of ART on anaemia during HIV infection.

o Line 49 – refer to point on line 26-27.

o Line 83-87: see comment above regarding iron deficiency definition and reference to the second manuscript; the prevalence of BM-ID in the subset of individuals assessed in the second manuscript was higher (48%) and could be cross-referenced here. Should Line 87 use ref 46, not 45? Reference 46 used BM iron rather than MCV as the definition.

o Line 92-94: Caution should be employed in making this inference – these patients are still all severely anaemic with multiple co-morbidities, and a high proportion are still immunosuppressed. It is possible the discrepancies are more likely related to assessment of iron status.

Formatting / Spelling:

• Abstract: repetition of “of anaemia” in Methods paragraph.

• Discussion: line 39: “therefore”

• Discussion: line 91: “severe”

• Discussion: line 92: “anaemia” not “anaemic"

• Discussion: line 94: “conditions such as iron deficiency are….”

Reviewer #2: i would like to thank the authors for submitting this article for submission. The article highlights the very important problem of severe anemia in HIV infected adults in Malawi and the multiple etiologies associated with this diagnosis. i believe this article contributes something new to the body of evidence already available for decision makers. I would like to recommend this article for publication but first I would like to authors to address a few concerns.

Major comments

in the methods section on page 10, the authors note that all HIV infected patients over 18 years old admitted to the general medical ward with severe anemia were approached and recruited for the study if they provided informed consent. How was the sample size (n=199) determined? this was not explained in the methods section.

How did the authors come about the 15 initial etiological factors that were explored for possible associations with severe anemia in HIV infected adults? some of the factors listed such as malaria, zidovudine use, iron deficiency have long been associated with anemia but others such as EBV, CMV, vitamin B12 not so much. Could the authors expand on this further in the introduction, discussion?

More than half of the study population was not on ART with low CD4 counts and high viral loads- thus this study population may not be generalizable to other adult HIV-infected populations in resource poor countries as the majority of the population were not on standard ART treatment and had a higher risk of mortality. the study would be more informative if the outcomes were looked at separately in patients on ART and those not on ART to determine if there was a significant difference in mortality, as well as in the associations with the different etiological factors.

Minor comments:

Figure 1: the title and explanation of the figure is in the manuscript but not on the figure itself. Similar comment for Figure 2, readers should be able to understand the figures by looking at them without having to refer to the manuscript for detailed explanation.

in figure 3, what do the plotted numbers represent and how will the readers interpret them? are they median values with interquartile ranges or hazard ratios with 95% confidence intervals? this was explained in the manuscript but should also be clear from looking at the figure itself.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2020 Feb 25;15(2):e0218695. doi: 10.1371/journal.pone.0218695.r002

Author response to Decision Letter 0

2 Oct 2019

To: PLOS ONE Kostas Pantopoulos, PhD Academic Editor

Subject: Rebuttal letter resubmission Manuscript ID: PONE-D-19-15827

September 2019

Dear Kostas Pantopoulos PLOS ONE,

We thank you for considering our manuscript titled; “Severe anaemia complicating HIV in Malawi; multiple co-existing aetiologies are associated with high mortality” (PONE-D-19-15827) for publication in your journal. We would like to highlight that we initially submitted two manuscripts back to back. The other manuscript titled; “Hepcidin and conventional markers to detect iron deficiency in severely anaemic HIV-infected patients in Malawi” (PONE-D-19-15824) has also been resubmitted. Please receive the suggested changes and answers considering the issues raised by the reviewer. In our reply we indicated the original text (with respective line and page numbers) and changes are underlined. As requested we have tracked change all changes made in the original manuscript.

Major comments:

General. Data availability:

1. All data was presented in the manuscript wit the exception the notification in line 393 (page 15): “TB medication itself was not associated with the severity of anaemia (data not shown).”

• We adjusted this and provided the data directly. TB medication itself was not associated with the severity of anaemia (OR 1.6 95%CI 0.7-3.8).

• All data was given within the manuscript as described above. We did not explicitly ask for medical ethical permission to publish our database online. In case individual readers are interested in using anonymised data we can and will share our database or requested data, we have added this to the results section

• Methods were adjusted line 214 (page 7; Additional raw data can be requested by contacting the corresponding author.

Ad reviewer 1:

Introduction

2. It would be useful to include brief consideration of how important anaemia is clinically in relation to other co-morbidities (i.e. how important is correction of anaemia per se).

• We have added the requested item to the introduction. Line 63(Page3). Anaemia treatment may even improve survival (1).

3. The authors should specifically mention inflammation as a potential cause of anaemia (this will likely be present during many of the infections, but may not be restricted to these).

• We agree with the reviewer. Line 66-68 (page3) introduction were rephrased: Severe anaemia in HIV infection has been associated with micronutrient deficiencies, infections (viral, bacterial and parasitic) and inflammation, medication induced (Zidovudine and co-trimoxazole), neoplastic diseases (7-11).

Methods

4. The authors confirm whether blinding was used in bone marrow analyses.

• We confirm that blinding was used. The method section has been adjusted: line 159 (page 5): Bone marrow analysis was performed by examiners that were blinded to patient data.

5. Methods/Results/Discussion (line 83): iron deficiency – the parallel manuscript demonstrates that MCV is a poor index of iron status (bone marrow iron staining, considered gold standard). The authors should make this clear in the present study in the text / table footnotes (e.g. by referring to the second manuscript) – it is therefore unclear how useful this part of the analysis is. Could the authors incorporate bone marrow iron status into analysis, for the subset of patients for whom that information is available?

• We specifically chose the MCV as it a) was the best (suboptimal) conventional marker in our setting b) was available in more patients than the subgroup of patients with a bone marrow and c) is available in most African settings and therefor applicable to the settings most interested readers work in. We have added the requested data however to the footnote of table 1.

• Methode line 178 (page 6): In a sub study (Huibers et al PLOSone2019) we evaluated bone marrow iron deficiency toward several conventional blood makers; MCV (fl), MCH (pg/cells), Fe (umol/l), Ferritine (ug/dl), TFr1 receptor (nmol/l), TrF index (stFR/Log ferritine). All markers showed suboptimal correlations AUCROC < 0.6 with bone-marrow iron deficiency. Of the conventional markers MCV performed best; AUCROC of 0.545. MCV was used to identified iron deficiency as it was the best (suboptimal) conventional marker in our setting, was available in more patients than the just the subgroup (n=76) of patients with a bone marrow result plus is available and known in most African settings (3, 22, 30).

• Footnote Table 2. Distribution and multivariate analysis of co-existing factors associated with severe (Hb≤ 70 g/l- (Hb> 50 g/l) versus very severe anaemia (Hb≤ 50 g/l) in HIV-infected adults in Malawi. 1 A total of 28-blood cultures were positive, the most common organisms were E. coli (42.9%; 12/28) and non-Typhoid Salmonella (17.9 %; 5/28). 2 Iron deficiency was defined by MCV≤ 83 fl; In a sub study (Huibers et al PLOSone2019) we evaluated bone marrow iron deficiency toward several conventional blood makers; MCV (fl), MCH (pg/cells), Fe (umol/l), Ferritine (ug/dl), TFr1 receptor (nmol/l), TrF index (stFR/Log ferritine). All markers showed suboptimal correlations AUCROC < 0.6 with bone-marrow iron deficiency. Of the conventional markers MCV performed best; AUCROC of 0.545.

• Line 449 (page 16) discussion was rephrased: Iron deficiency is complicated by a lack of peripheral iron markers and diagnosing remains challenging as peripheral blood markers as MCV poorly reflect bone-marrow iron deficiency especially in the context of HIV-infection (Huibers et al PLOSone 2019).

Results

6. Table 2: - % with severe anaemia (in title row) should be corrected.

• This has been adjusted (line 388 page 11): Overall N=199(100%) and severe anaemia N=115/199 (57.8%)

7. Table 2: In the table footer, the authors should clarify that analyses are comparing severe with very severe anaemia.

• Rephrased (line 398 page 12): Table 2. Distribution and multivariate analysis of co-existing factors associated with severe (Hb≤ 70 g/l- Hb> 50 g/l) versus very severe anaemia (Hb≤ 50 g/l) in HIV-infected adults in Malawi.

8. The methods section states that variables with P<0.1 in univariate are included in multivariate analysis: it appears that variables with P>0.1 are included – this should be clarified / corrected.

• We have accidentally added a VL > 1000 copies with a p-value of 0.163 in the multivariate analysis as well. Wherefore excuse. We adjusted table 2 accordingly. Multivariate outcome did not change significantly (line 398 page 12).

9. Renal function: normal – is it intentional that no statistics are included here?

• In the Logistic regression impaired function and end stage renal function were compared toward a normal renal function. We agree with the reviewer that we displayed this in a confusing way. To display the results more clearly we deleted the normal renal function from table two (line 398 page 12).

10. Do particular co-morbidities frequently occur together?

• There is obvious overlap of these variables as is graphically displayed in the new supplementary figure S1. Given that this overlap and the wide spread, we have chosen not to over interpretate our dataset and have restricted our analysis to simple uni- and multivariate analyses.

• The results section was adjusted line 247 (page9) Overlap of risk factors is visually displayed in supplemental figure S1.

11. Given the pattern of mortality (approx. 50% within 1 year, approx. 80% of whom died within 60 days). The analysis could be refined to compare patients who died within 60 days vs those who died later and/or those who survived: was there a difference in anaemia group, haemoglobin concentration, or co-morbidities between these groups?

• We appreciate this question from the reviewer. To display this information we created a supplementary table S1. The prevalence of EBV was significant different between the group of patients who died within 60 days (50%) and after 60 days (12.5%); p-value 0.006. However when we do a multivariate cox regression for mortality EBV is not significant related toward mortality either within 60 as after 60 days. We therefore did not wanted to include the outcome of this table directly in the results section. But with regard to the question we added the supplemental table S1. Results of the uni and multivariate analysis are aswell provided in supplementary table S2. See comment below.

• Methods were adjusted in line 204 (page 6). Moreover a sub analysis was performed to evaluate the distribution of co-existing factors between mortality within and after 60 days. Group comparisons for categorical data were performed using the χ2 test or Fisher’s exact test, and for continuous data using the t-test or the Wilcoxon rank-sum test.

• Results were adjusted in line 302 (page 10): Distribution of co-existing factors between mortality within 60 days and overall mortality in HIV-infected adults in Malawi was given in supplementary table S1

• See for rephrasing on table S2 point 12.

12. Do any particular combinations associate with anaemia severity, haemoglobin concentration or mortality/mortality within first 60 days?

• We included outcome in the result section in an additional supplemental figure and results were rephrased (Line 306 page 10): Folate deficiency and end stage renal disease were associated with overall mortality with Hazard Ratio 2.2 (95% CI 1.2-3.8) and Hazard Ratio 3.2 (95% CI 1.6-6.2) respectively (figure 3). End stage renal disease (Hazard Ratio 2.7 95% CI 1.2-6.2) was associated with mortality within 60 days, supplementary table S2.

• Secondly this sub-analysis was added to the method Methods were adjusted in line 207 page: Uni- and multivariate analyses were done using logistic regression and Cox regression to describe predictors of overall mortality. A sub- analysis was performed for mortality within 60 days.

13. Table 1 / Text: there is a discrepancy between the mortality reported in the text and in the Table. This should be corrected.

• This was corrected. Line 300 page 10: During the one-year follow-up period, 101 study patients (50.8%) died. The median time to death was 17.5 days (IQR 6-55) and 81 (80.2%) of these deaths occurred within 60 days of admission (figure 2).

14. Results: Figure 3 – the position of the “zero” line should be aligned with 0.

• We have adjusted the lines to 1.0 (the neutral OR)

• With adjusting the neutral OR we found a minor error, which we corrected. The text described a minimally different outcome, difference in decimal (HR/CI) than shown in figure 3. This has been adjusted. Overall outcome did not change.

• Folate deficiency (≤3 ng / l); HR 2.1 95% CI 1.2-3.6 → HR 2.0 (95% CI 1.2-3.6) and stage renal disease (GFR ≤15); HR 3.2 5% CI 1.6-6.2) → HR 3.0 5% CI 1.5-5.9.

• Line 303 page 10: Folate deficiency and end stage renal disease were associated with overall mortality with Hazard Ratio 2.0 (95% CI 1.2-3.6) and Hazard Ratio 3.0 (95% CI 1.5-5.9) respectively (figure 3).

• Figure 3. Risk factors for 365-day mortality in HIV-infected patients with severe anaemia. Univariate and multivariate Cox regression outcome (Hazard Ratios 95% CI). Folate defiency (≤3 ng/l) HR 2.0 95% CI 1.2-3.6 and end stage renal disease (GFR ≤15); HR 3.0 5% CI 1.5-5.9, were associated with overall mortality.

Discussion

The discussion is long and could be made more concise. In places it seems to infer a causative relationship between anaemia and mortality, but the present analyses do not allow such inferences to be made; HIV pathogenesis is complex, and although anaemia may contribute to morbidity and mortality, the data show that there are many other co-morbidities that could also explain the mortality risk.

15. We have rephrased several section of the discussion to comply with these comments:

• Discussion line 455, page 14: Severe anaemia in HIV-infected patients in a resource limited setting, such as Malawi, is a multi-causal critical condition associated with high mortality.

• Discussion line 459, page 14 rephrased: However HIV pathogenesis is complex and although anaemia may contribute to morbidity and mortality, the data suggest that the potential aetiology is multi-causal. Previous studies reported an estimated one-year mortality of to 30-55% in severe anaemic versus 3.7% in non-anaemic HIV-infected patients in resource limited settings (5, 6)

• Conclusion underlines the opinion of the reviewer: Line 642, page 17: Severe anaemia in HIV-infected patients is therefore a critical indicator of mortality and requires urgent and multiple interventions. “ That we emphasise that criticality of anaemia without phrasing the direct and only contributor of mortality in the complex HIV-pathophysiology.

16. In general the discussion can be made more concise, e.g. see repetition in lines 10-11 and 20-21. We have made several adjustments. All adjustments are explained and displayed in the comments below.

• Discussion line 457, page 14: Anaemia in HIV-infected patients is descripted as an independent predictor of mortality with a direct effect of decreasing haemoglobin concentrations.

• Discussion line 463, page 14: The seriousness of severe anaemia and its complexity in HIV-infected patients needs to be better recognized.

17. Lines 14-15: this could be interpreted as anaemia being causal for mortality; it is part of a complex disease state. Figure 3 found no association of anaemia severity with mortality risk (noting that all individuals had at least severe anaemia). However, in concert with much previous literature, the data are consistent with an associations between severe anaemia and mortality in HIV infection given this mortality rate. Suggest revising to “a multi-causal critical condition associated with high mortality”.

• See comment 15, rephrasing Discussion line 455, page 14.

18. Line 26-27: it is a little unclear whether the authors are implying that severe anaemia is itself a prime cause of mortality (that therefore should be addressed per se), or rather whether it is a clear marker of severe disease (line 30) that would likely be improved through addressing co-existing conditions associated with advanced HIV (infections etc). The literature cited regarding anaemia and mortality is correlative, so the latter interpretation is favoured in my view.

• We appreciated this comments and adjusted the discussion, line 334, page14 is rephrased: The seriousness of severe anaemia and its complexity in HIV-infected patients needs to be better recognized. Irrespective if severe anaemia is the direct cause of the mortality or merely linked to other contributors causing both death and anaemia, one can state that severe anaemia should trigger clinicians as an emergency sign and therefor should be grouped as a stage 4 condition in stead of the current stage 3 classification (33).

19. Line 34: add “and inflammation”?

• We agree with the reviewer and adjusted discussion line 471, page 14: HIV may cause anaemia directly through an inhibitory effect of the HIV-virus on the erythropoietin progenitor cells in the bone marrow, or indirectly through opportunistic infections and/or inflammation causing anaemia (37).

20. Line 36: the lack of effect here could reflect all patients having Hb<7g/dL in the present study.

• We agree with the reviewer and adjusted discussion line 474, page 14: In our cohort this was not confirmed however our cohort consist with patients who all experienced severe anaemia.

21. Line 39-41: the authors could cite relevant data on the impact of ART on anaemia during HIV infection.

• Discussion Line 555, page 15 was adjusted: Therefore they are likely to continue having high levels of life-threatening anaemia, as the complex multifactorial ethology of the severe anaemia might be marker of sever HIV disease. At last the effect of medication as Zidovidine and trimoxazole on the bone marrow should also be taken in consideration. However possible due to the bias of all severe anaemic HIV-infected patients, this was not confirmed as directly effective on severity of anaemia.

22. Line 49 – refer to point on line 26-27.

• We appreciate this suggestion. Discussion line 553, page 15 is rephrased; Consequently they are likely to continue having high levels of life-threatening anaemia, as the complex multifactorial ethology of the severe anaemia might be marker of sever HIV disease.

23. Line 83-87: see comment above regarding iron deficiency definition and reference to the second manuscript; the prevalence of BM-ID in the subset of individuals assessed in the second manuscript was higher (48%) and could be cross-referenced here. Should Line 87 use ref 46, not 45? Reference 46 used BM iron rather than MCV as the definition.

• See point 5 Line 449 (page 16) discussion was rephrased: Iron deficiency is complicated by a lack of peripheral iron markers and diagnosing remains challenging as peripheral blood markers as MCV poorly reflect bone-marrow iron deficiency especially in the context of HIV-infection (Huibers et al PLOSone 2019).

• We agree with the reviewer. Reference 46 was replaced by 45

24. Line 92-94: Caution should be employed in making this inference – these patients are still all severely anaemic with multiple co-morbidities, and a high proportion are still immunosuppressed. It is possible the discrepancies are more likely related to assessment of iron status

• Discussion line 615-616, page 16 was rephrased: In this group of HIV-infected patients with better immune systems the aetiology of severe anaemia may be more similar to the multifactorial aetiology of non-HIV infected patients. Conditions such as iron deficiency are more alike the HIV-uninfected groups.

Formatting / Spelling:

25. Abstract: repetition of “of anaemia” in Methods paragraph.

• Rephrased: Fifteen potential causes and associations with anaemia severity and mortality were explored.

26. Discussion: line 39: “therefore”

• This line was already rephrased with regard to above comments. Discussion line 525 page 15: Consequently they are likely to continue having high levels of life-threatening anaemia, as the complex multifactorial ethology of the severe anaemia might be marker of sever HIV disease.

27. Discussion: line 91: “severe” | line 92: “anaemia” not “anaemic" line 94: “conditions such as iron deficiency are….”

• Rephrased discussion line 587-588, page 16 “In this group of HIV-infected patients with better immune systems the aetiology of severe anaemia may be more similar to the aetiology of non-HIV infected patients. Conditions such as iron deficiency are more alike the HIV-uninfected groups.

Reviewer 2;

Major comments

28. Sample size: in the methods section on page 10, the authors note that all HIV infected patients over 18 years old admitted to the general medical ward with severe anemia were approached and recruited for the study if they provided informed consent. How was the sample size (n=199) determined? this was not explained in the methods section.

• Methods- statistics were adjusted. Methode line 193 (page 6). The study was primarily designed to give a complete overview of the potential factors associated with severe anaemia with a special attention of the role of lymphomas in a severe anemic african patients ipopulation infected with HIV. A sample size of 200 would be able to detect an estimated prevalence of 5 or 10% with confidence intervals of 2.4% - 9% and 6- 15% respectively and was therefor chosen.

29. How did the authors come about the 15 initial etiological factors that were explored for possible associations with severe anemia in HIV infected adults? some of the factors listed such as malaria, zidovudine use, iron deficiency have long been associated with anemia but others such as EBV, CMV, vitamin B12 not so much. Could the authors expand on this further in the introduction, discussion?

• The 15 potential factors studied were based on two previous studies performed in Malawi and a systemic review on this topic:

• Calis JC, Phiri KS, Faragher EB, Brabin BJ, Bates I, Cuevas LE, et al. Severe anemia in Malawian children. N Engl J Med. 2008;358(9):888-99.

• Lewis DK, Whitty CJ, Walsh AL, Epino H, Broek NR, Letsky EA, et al. Treatable factors associated with severe anaemia in adults admitted to medical wards in Blantyre, Malawi, an area of high HIV seroprevalence. Transactions of the Royal Society of Tropical Medicine and Hygiene. 2005;99(8):561-7.

• Calis JC, van Hensbroek MB, de Haan RJ, Moons P, Brabin BJ, Bates I. HIV-associated anemia in children: a systematic review from a global perspective. AIDS. 2008 Jun 19;22(10):1099-112.

• This comment was rephrased in the method section line 168 (page 5): A total of 15 potential factors involved in the aetiology of severe anaemia were investigated, factors were based on two previous studies performed in Malawi and a systemic review on this topic (30-32).

30. More than half of the study population was not on ART with low CD4 counts and high viral loads- thus this study population may not be generalizable to other adult HIV-infected populations in resource poor countries as the majority of the population were not on standard ART treatment and had a higher risk of mortality. The study would be more informative if the outcomes were looked at separately in patients on ART and those not on ART to determine if there was a significant difference in mortality, as well as in the associations with the different etiological factors.

• We agree that it appears informative to stratify our analyses for ART use at enrolment, which applied to nearly half the study population. In both the mortality (overall and 60 days) and the severity analyses ART use at enrolment was not associated with these outcomes. Therefore we decided not to present a stratified analysis in this paper. In line 403 of the discussion we further highlight the fact that the findings of our study remain very relevant because many HIV-infected patients in resource-limited settings present late in the course of their disease or are unable to access reliable supplies of ART. This may especially apply to those with severe anaemia, as it is a complication of advanced disease.

Minor comments:

31. Figure 1: the title and explanation of the figure is in the manuscript but not on the figure itself.

• Figure 1. Total number aetiologies for severe anaemia co-existing in each patient (n=199). Mean is 3 factors (SD 1.3), range 1-8. Aetiologies for severe anaemia include: 1) Unsuppressed HIV-infection; viral load ≥1000 copies/ml. 2) TB: one or more of the following were present: a) positive sputum culture, b) chest X-ray with signs of pulmonary TB and/or c) on going TB treatment at time of enrolment d) clinical diagnosis by local doctor including unknown generalized lymphadenopathy and/or night sweats of > 30 days and of unknown origin e) caseating granulomata in the bone marrow trephine. 3) Malaria: presence of malaria parasites in a thick blood film. 4) Parvovirus B19: viral load of >1000 copies/ml. 5) Cytomegalovirus (CMV); load of >100 copies/ml. 6) Epstein-Barr virus (EBV); viral load >100 copies/ml. 7) Bacteraemia; a blood culture growing a potential pathogen. 8) Underweight (BMI ≤18.5). 9) Serum folate deficiency (≤3 ng/l). 10) Vitamin B12 deficiency (≤180 pg/ml). 11). Iron deficiency was defined as MCV≤ 83 fl (3, 22, 33). 12) Zidovudine usage. 13) Co-trimoxazole usage. 14) Bone marrow disorders; lympho-proliferative disease, myeloid-proliferative disease or MDS. 15) Renal impairment: a GFR which either indicated impaired (GFR 15–59 ml/min/1.73 m2) or End Stage (GFR ≤15 ml/min/1.73 m2) Renal Disease (21, 34).

32. Similar comment for Figure 2, readers should be able to understand the figures by looking at them without having to refer to the manuscript for detailed explanation.

• Line 385 page 10; Figure 2. Kaplan Meyer survival curve over time (days) for adult Malawian patients with HIV infection and severe anaemia during 365 days follow-up. Abbreviation: 95% confidence interval (95%CI).

33. Figure 3, what do the plotted numbers represent and how will the readers interpret them? are they median values with interquartile ranges or hazard ratios with 95% confidence intervals?

• Line 388 page 10; Rephrased: Figure 3. Risk factors for 365-day mortality in HIV-infected patients with severe anaemia. Univariate and multivariate Cox regression outcome (Hazard Ratios 95% CI).

We hope to have clarified outstanding questions and improved the manuscript according to the concerns raised. Please do not hesitate to contact us if you have any further questions.

With many thanks for your consideration, and on behalf of all the authors.

Minke Huibers, MD

Attachment

Submitted filename: 1. Response to Reviewers manuscript I.docx

Click here for additional data file. (74.6KB, docx)

Decision Letter 1

Kostas Pantopoulos

23 Oct 2019

PONE-D-19-15827R1

Severe anaemia complicating HIV in Malawi; multiple co-existing aetiologies are associated with high mortality.

PLOS ONE

Dear mrs Huibers,

Thank you for submitting your manuscript to PLOS ONE. Both reviewers found that the revised manuscript is improved and has addressed most of their concerns. However, reviewer 1 raised a few remaining minor issues, which require attention. Therefore, we invite you to address them in a revised version of the manuscript.

We would appreciate receiving your revised manuscript by Dec 07 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Kostas Pantopoulos, PhD

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Thank you to the authors for their engagement with the reviewer comments. They have made several changes and consequently the manuscript is improved. A few minor points remain.

1. There is a discrepancy between the median days to death reported in the abstract, versus that reported in Table 1 / line 283 – this should be corrected.

2. Point 2 from the previous review (“Introduction (paragraph 2): it would be useful to include brief consideration of how important anaemia is clinically in relation to other co-morbidities (i.e. how important is correction of anaemia per se)”).

• The authors have added a sentence referring to reference 1 (Mocroft et al, 1999) stating “Anaemia treatment may even improve survival”. The data presented in the large Mocroft study do not address treatment of anaemia and indeed they conclude (final sentence) that “Further follow-up is needed to determine…. whether an increase in haemoglobin decreases the risk of death”.

• This therefore does not seem an appropriate reference to support the claim – this should be revised or toned down, or appropriate support presented from other sources.

3. Iron deficiency based on MCV:

• I accept that this is now clarified in the recent manuscript. However, I still feel that classifying these individuals as “iron deficient” when the test performance is so poor (AUC=0.54) could be misleading, even though this is the highest (least low) AUC among the markers tested.

• In my opinion, it would therefore be more appropriate to classify explicitly as “low MCV (<83 fL)” or similar (rather than “iron deficiency”), indicating e.g. in the text that this definition is used elsewhere as an index of iron deficiency, but noting that it poorly identified bone marrow iron deficiency in this group (Huibers et al, 2019 – other manuscript).

4. General point: attention needs to be given to manuscript organisation and spelling/grammar in places – a check should be made. For example:

• Remove repetition in discussion: see lines 399-402 and lines 411-414.

• Table S1 / S2: remove repetition in the figure legends (text is duplicated).

• Line 356: sever

• Line 173 and elsewhere: ferritine – change to ferritin

Reviewer #2: Authors have addressed all my concerns. Minor revisions of spelling and grammar recommended in discussion section:

Discussion, line 326: ‘Anaemia in HIV-infected patients is descripted as an independent predictor’ should be corrected to ‘Anaemia in HIV-infected patients is an independent predictor’

Discussion, line 330: ‘reported an estimated one-year mortality of 30 - 55% in severe anaemic vs 3.7%’ should read ‘reported an estimated one-year mortality of 30 - 55% in severely anaemic vs 3.7%’

Discussion, line 344: ‘however our cohort consist with’ should read ‘however our cohort consisted of’

Discussion, line 356: ‘ the complex multifactorial ethology of the severe anaemia might be marker of sever HIV disease’ should read ‘ the complex multifactorial etiology of the severe anaemia might be a marker of severe HIV disease’

Discussion, line 357: 'At last the effect of medication as Zidovudine' should read ‘Lastly, the effects of medications such as Zidovudine’

Discussion, line 358: ‘ However possible due to the bias’ should read ‘however, possibly due to the bias’

Discussion, line 394 and 398- repetition of the statement ‘the prevalence of iron deficiency in our study is higher than…'

Discussion, line 408: 'sever immune suppression' should read 'severe immune suppression’

Discussion, line 413: should read ‘peripheral blood markers such as MCV poorly reflect bone-marrow iron deficiency'

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2020 Feb 25;15(2):e0218695. doi: 10.1371/journal.pone.0218695.r004

Author response to Decision Letter 1

15 Jan 2020

To: PLOS ONE Kostas Pantopoulos, PhD Academic Editor

Subject: Rebuttal letter resubmission 2 Manuscript ID: PONE-D-19-15827

January 2020

Dear Kostas Pantopoulos PLOS ONE ,

We thank you for accepting our manuscripts titled; “Severe anaemia complicating HIV in Malawi; multiple co-existing aetiologies are associated with high mortality” (PONE-D-19-15827) with minor changes for publication in your journal. We would like to highlight that we initially submitted two manuscripts back to back. The other manuscript titled; “Hepcidin and conventional markers to detect iron deficiency in severely anaemic HIV-infected patients in Malawi” (PONE-D-19-15824) has also been resubmitted with the requested minor changes. Please see below for our responses to the issues raised by the reviewer. We have indicated the original text (with respective line and page numbers) and changes are underlined. As requested we have tracked all changes made in the original manuscript.

It was requested that the raw data should be available as supplementary information for further review, in line with Plos One policy. We are very willing to do so and submitted therefore an anonymised database as a supplementary file.

Reviewer 1:

1. There is a discrepancy between the median days to death reported in the abstract, versus that reported in Table 1 / line 283 – this should be corrected.

• We have adjusted the abstract. Line 46 (Page2): Overall mortality was high (53%; 100/199) with a median time to death of 17.5 days (IQR 6-55) days.

2. Point 2 from the previous review (“Introduction (paragraph 2): it would be useful to include brief consideration of how important anaemia is clinically in relation to other co-morbidities (i.e. how important is correction of anaemia per se)”).

• The authors have added a sentence referring to reference 1 (Mocroft et al, 1999) stating “Anaemia treatment may even improve survival”. The data presented in the large Mocroft study do not address treatment of anaemia and indeed they conclude (final sentence) that “Further follow-up is needed to determine…. whether an increase in haemoglobin decreases the risk of death”.

• This therefore does not seem an appropriate reference to support the claim – this should be revised or toned down, or appropriate support presented from other sources.

• We agree and have changed the reference: Associations of anemia, treatments for anemia, and survival in patients with human immunodeficiency virus infection. Sullivan : J Infect Dis. 2002 May 15;185 Suppl 2:S138-42.

• Line 76 page 3: Anaemia treatment may even improve survival (7).

3. Iron deficiency based on MCV:

• I accept that this is now clarified in the recent manuscript. However, I still feel that classifying these individuals as “iron deficient” when the test performance is so poor (AUC=0.54) could be misleading, even though this is the highest (least low) AUC among the markers tested.

• In my opinion, it would therefore be more appropriate to classify explicitly as “low MCV (<83 fL)” or similar (rather than “iron deficiency”), indicating e.g. in the text that this definition is used elsewhere as an index of iron deficiency, but noting that it poorly identified bone marrow iron deficiency in this group (Huibers et al, 2019 – other manuscript).

• We appreciate the concerns of the reviewer. Given the use of MCV as marker for iron deficiency in previous reports and the MCV evaluation and explanation given in our sub study, we have taken up the reviewer’s helpful suggestion and have now used MCV ≤ 83 fl as the definition of iron deficiency throughout.

• Lines 242 page 6 of the methods section were adjusted: In a sub study (Huibers et al PLOS One 2019) we evaluated bone marrow (BM) iron deficiency using several conventional blood markers; MCV (fl), MCH (pg/cells), Fe (umol/l), ferritin (ug/dl), TFr1 receptor (nmol/l), TrF index (stFR/Log ferritin). All markers showed suboptimal correlations (i.e. AUCROC < 0.6) with BM iron deficiency, though MCV performed best (AUCROC 0.545). MCV was therefore used to identified iron deficiency as it was the best (though suboptimal) conventional marker in our setting. It was also available for patients beyond the subgroup (n=76) who had a bone marrow result and it is available in most African settings (3, 22, 33). Since an MCV <83fL is commonly used in other studies and guidelines (3, 22, 33) and it was the best predictor of BM iron deficiency in our setting, we used it as the marker for iron deficiency in this analysis. As the AUC-ROC was suboptimal we chose to call it MCV ≤83fL rather than iron deficiency.

• Iron deficiency was therefore replaced by MCV ≤ 83 fl throughout the analysis/discussion

o Figure 3, Table 2, Table S1 and S2, line 408, 387, 390 (page 15) and line 240 (page 8).

3. General point: attention needs to be given to manuscript organisation and spelling/grammar in places – a check should be made. For example:

• Remove repetition in discussion: see lines 399-402 and lines 411-414.

• Table S1 / S2: remove repetition in the figure legends (text is duplicated).

• Line 356: sever

• Line 173 and elsewhere: ferritine – change to ferritin

• We have adjusted the spelling/grammar accordingly; moreover we have as suggested by the reviewer reviewed spelling and grammar. The sections have been adjusted as requested and the manuscript has been revised by a native English speaker.

Reviewer 2:

4. Discussion, line 326: ‘Anaemia in HIV-infected patients is descripted as an independent predictor’ should be corrected to ‘Anaemia in HIV-infected patients is an independent predictor’

5. Discussion, line 330: ‘reported an estimated one-year mortality of 30 - 55% in severe anaemic vs 3.7%’ should read ‘reported an estimated one-year mortality of 30 - 55% in severely anaemic vs 3.7%’

6. Discussion, line 344: ‘however our cohort consist with’ should read ‘however our cohort consisted of’

7. Discussion, line 356: ‘ the complex multifactorial ethology of the severe anaemia might be marker of sever HIV disease’ should read ‘ the complex multifactorial etiology of the severe anaemia might be a marker of severe HIV disease’

8. Discussion, line 357: 'At last the effect of medication as Zidovudine' should read ‘Lastly, the effects of medications such as Zidovudine’

9. Discussion, line 358: ‘ However possible due to the bias’ should read ‘however, possibly due to the bias’

10. Discussion, line 394 and 398- repetition of the statement ‘the prevalence of iron deficiency in our study is higher than…'

11. Discussion, line 408: 'sever immune suppression' should read 'severe immune suppression’

12. Discussion, line 413: should read ‘peripheral blood markers such as MCV poorly reflect bone-marrow iron deficiency'

• The sections have been adjusted as requested and the manuscript has been revised by a native English speaker.

We hope our revisions in response to the reviewers’ comments have improved the manuscript and please do not hesitate to contact us if you have any further questions.

With many thanks for your consideration, on behalf of all the authors.

Minke Huibers, MD, PhD.

Decision Letter 2

Kostas Pantopoulos

23 Jan 2020

Severe anaemia complicating HIV in Malawi; multiple co-existing aetiologies are associated with high mortality

PONE-D-19-15827R2

Dear Dr. Huibers,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Kostas Pantopoulos, PhD

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Acceptance letter

Kostas Pantopoulos

12 Feb 2020

PONE-D-19-15827R2

Severe anaemia complicating HIV in Malawi; multiple co-existing aetiologies are associated with high mortality

Dear Dr. Huibers:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Kostas Pantopoulos

Academic Editor

PLOS ONE

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 Data

    (DTA)

    Click here for additional data file. (92.3KB, dta)
    S1 Fig

    (TIF)

    Click here for additional data file. (1.8MB, tif)
    S1 Table

    (DOCX)

    Click here for additional data file. (94KB, docx)
    S2 Table

    (DOCX)

    Click here for additional data file. (85.9KB, docx)
    Attachment

    Submitted filename: 1. Response to Reviewers manuscript I.docx

    Click here for additional data file. (74.6KB, docx)

    Data Availability Statement

    All relevant data are within the manuscript and its Supporting Information files.

    Articles from PLoS ONE are provided here courtesy of PLOS

    RESOURCES