• View in gallery

    Diagnostic accuracy of peripheral blood microscopy compared with invasive tissue aspirate microscopy, according to (A) parasite concentration method (B) and human immunodeficiency virus (HIV) status. FN = false negative; FP = false positive; HIV-neg = HIV-seronegative patients; HIV-pos = HIV-seropositive patients; PBMC = peripheral blood mononuclear cell; TN = true negative; TP = true positive. Numbers in brackets and bars represent 95% confidence interval (CI). Combined PBMC/buffy denotes the cumulative result from smears of both PBMC and the buffy coat of a given patient.

  • View in gallery

    Overlap of results from microscopic examination of tissue aspirates, PBMC isolation, and buffy coat. Note that buffy-coat smears from 4/301 patients during the study were either lost or not examined, yielding a total of 297 patients depicted. Percentages in the circles are of the total population. PBMC = peripheral blood mononuclear cells.

  • View in gallery

    Leishmania donovani amastigotes in the (A and B) spleen, (C) PBMC, and (D) buffy coat of visceral leishmaniasis patients, using Giemsa stain. Arrowheads indicate amastigotes with visible nucleus and kinetoplast.

  • 1.

    Alvar J, Velez ID, Bern C, Herrero M, Desjeux P, Cano J, Jannin J, den Boer M, 2012. Leishmaniasis worldwide and global estimates of its incidence. PLoS One 7: e35671.

    • Search Google Scholar
    • Export Citation
  • 2.

    van Griensven J, Diro E, 2012. Visceral leishmaniasis. Infect Dis Clin North Am 26: 309322.

  • 3.

    Alvar J, Bashaye S, Argaw D, Cruz I, Aparicio P, Kassa A, Orfanos G, Parreno F, Babaniyi O, Gudeta N, Canavate C, Bern C, 2007. Kala-azar outbreak in Libo Kemkem, Ethiopia: epidemiologic and parasitologic assessment. Am J Trop Med Hyg 77: 275282.

    • Search Google Scholar
    • Export Citation
  • 4.

    Bern C, Maguire JH, Alvar J, 2008. Complexities of assessing the disease burden attributable to leishmaniasis. PLoS Negl Trop Dis 2: e313.

  • 5.

    Boelaert M, Verdonck K, Menten J, Sunyoto T, van Griensven J, Chappuis F, Rijal S, 2014. Rapid tests for the diagnosis of visceral leishmaniasis in patients with suspected disease. Cochrane Database Syst Rev 6: CD009135.

    • Search Google Scholar
    • Export Citation
  • 6.

    Cunningham J, Hasker E, Das P, El Safi S, Goto H, Mondal D, Mbuchi M, Mukhtar M, Rabello A, Rijal S, Sundar S, Wasunna M, Adams E, Menten J, Peeling R, Boelaert M, 2012. A global comparative evaluation of commercial immunochromatographic rapid diagnostic tests for visceral leishmaniasis. Clin Infect Dis 55: 13121319.

    • Search Google Scholar
    • Export Citation
  • 7.

    Medrano FJ, Canavate C, Leal M, Rey C, Lissen E, Alvar J, 1998. The role of serology in the diagnosis and prognosis of visceral leishmaniasis in patients coinfected with human immunodeficiency virus type-1. Am J Trop Med Hyg 59: 155162.

    • Search Google Scholar
    • Export Citation
  • 8.

    ter Horst R, Collin SM, Ritmeijer K, Bogale A, Davidson RN, 2008. Concordant HIV infection and visceral leishmaniasis in Ethiopia: the influence of antiretroviral treatment and other factors on outcome. Clin Infect Dis 46: 17021709.

    • Search Google Scholar
    • Export Citation
  • 9.

    Diro E, Lynen L, Ritmeijer K, Boelaert M, Hailu A, van Griensven J, 2014. Visceral Leishmaniasis and HIV coinfection in east Africa. PLoS Negl Trop Dis 8: e2869.

    • Search Google Scholar
    • Export Citation
  • 10.

    Gidwani K, Picado A, Ostyn B, Singh SP, Kumar R, Khanal B, Lejon V, Chappuis F, Boelaert M, Sundar S, 2011. Persistence of Leishmania donovani antibodies in past visceral leishmaniasis cases in India. Clin Vaccine Immunol 18: 346348.

    • Search Google Scholar
    • Export Citation
  • 11.

    Boelaert M, el Safi S, Goetghebeur E, Gomes-Pereira S, Le Ray D, Van der Stuyft P, 1999. Latent class analysis permits unbiased estimates of the validity of DAT for the diagnosis of visceral leishmaniasis. Trop Med Int Health 4: 395401.

    • Search Google Scholar
    • Export Citation
  • 12.

    Chemli J, Abroug M, Fathallah A, Abroug S, Ben SM, Harbi A, 2006. Contribution of leukoconcentration in the diagnosis of Kala-azar in Tunisia [in French]. Med Mal Infect 36: 390392.

    • Search Google Scholar
    • Export Citation
  • 13.

    Deniau M, Canavate C, Faraut-Gambarelli F, Marty P, 2003. The biological diagnosis of leishmaniasis in HIV-infected patients. Ann Trop Med Parasitol 97 (Suppl 1): 115133.

    • Search Google Scholar
    • Export Citation
  • 14.

    Izri MA, Robineau M, Petithory JC, Rousset JJ, 1993. Visceral leishmaniasis. Parasitological diagnosis by leukocyte concentration [in French]. Presse Med 22: 1010.

    • Search Google Scholar
    • Export Citation
  • 15.

    Izri MA, Deniau M, Briere C, Rivollet D, Petithory JC, Houin R, Rousset JJ, 1996. Leishmaniasis in AIDS patients: results of leukocytoconcentration, a fast biological method of diagnosis. Bull World Health Organ 74: 9193.

    • Search Google Scholar
    • Export Citation
  • 16.

    Boelaert M, Bhattacharya S, Chappuis F, el Safi S, Hailu A, Mondal D, Rijal S, Sundar S, Wasunna M, Peeling RW, 2007. Evaluation of rapid diagnostic tests: visceral leishmaniasis. Nat Rev Microbiol 5: S30S39.

    • Search Google Scholar
    • Export Citation
  • 17.

    World Health Organization, 2010. Control of the Leishmaniases. Report of a meeting of the WHO Expert Committee on the Control of Leishmaniases, Geneva, March 22–26, 2010. Geneva, Switzerland: WHO Technical Report Series 949, World Health Organization.

    • Search Google Scholar
    • Export Citation
  • 18.

    Petithory JC, Ardoin F, Ash LR, Vandemeulebroucke E, Galeazzi G, Dufour M, Paugam A, 1997. Microscopic diagnosis of blood parasites following a cytoconcentration technique. Am J Trop Med Hyg 57: 637642.

    • Search Google Scholar
    • Export Citation
  • 19.

    Boelaert M, Rijal S, Regmi S, Singh R, Karki B, Jacquet D, Chappuis F, Campino L, Desjeux P, Le Ray D, Koirala S, Van der Stuyft P, 2004. A comparative study of the effectiveness of diagnostic tests for visceral leishmaniasis. Am J Trop Med Hyg 70: 7277.

    • Search Google Scholar
    • Export Citation
  • 20.

    Diro E, Lynen L, Mohammed R, Boelaert M, Hailu A, van Griensven J, 2014. High parasitological failure rate of visceral leishmaniasis to sodium stibogluconate among HIV co-infected adults in Ethiopia. PLoS Negl Trop Dis 8: e2875.

    • Search Google Scholar
    • Export Citation
  • 21.

    Green TA, Black CA, Johnson RE, 2001. In defense of discrepant analysis. J Clin Epidemiol 54: 210215.

  • 22.

    Lipman HB, Astles JR, 1998. Quantifying the bias associated with use of discrepant analysis. Clin Chem 44: 108115.

  • 23.

    Schachter J, 2001. In defense of discrepant analysis. J Clin Epidemiol 54: 211212.

 
 
 

 

 
 
 

 

 

 

 

 

 

Diagnosis of Visceral Leishmaniasis Using Peripheral Blood Microscopy in Ethiopia: A Prospective Phase-III Study of the Diagnostic Performance of Different Concentration Techniques Compared to Tissue Aspiration

View More View Less
  • 1 University of Gondar, Gondar, Ethiopia.
  • | 2 J.D. MacLean Centre for Tropical Diseases at McGill University, Montreal, Canada.
  • | 3 Institute of Tropical Medicine, Antwerp, Belgium.
  • | 4 School of Medicine, Addis Ababa University College of Health Sciences, Addis Ababa, Ethiopia.

Visceral leishmaniasis (VL) is a fatal parasitic disease. Unfortunately, diagnosis of VL in east Africa currently relies on aspiration of tissue from the spleen or bone marrow, which is painful and potentially dangerous. We sought to determine whether peripheral blood could be used instead of invasive tissue aspirates to diagnose VL, using three parasite concentration techniques. Three hundred and one consecutive people suspected of having VL were recruited. Compared with microscopy of tissue aspirates, the diagnostic accuracy of peripheral blood microscopy was as follows: whole blood thin smear sensitivity 1.5% (95% confidence interval [CI] 0.0–8.3) and specificity 100% (95% CI 76.8–100); buffy-coat smear sensitivity 19.5% (95% CI 14.3–25.6) and specificity 98.9% (95% CI 94.1–100); peripheral blood mononuclear cell (PBMC) smear sensitivity 33.7% (95% CI 27.3–40.5) and specificity 95.7% (95% CI 89.6–98.6). Sensitivity of PBMC smears was significantly higher in human immunodeficiency virus (HIV)-coinfected patients (N = 48/301); two-sample test of proportions, P = 0.0097; sensitivity 55.9% (95% CI 37.9–72.8) and specificity 92.9% (95% CI 66.1–99.8), and correlated with the degree of parasite load in the tissue. Combining the results from smears of both PBMC and buffy coat yielded a sensitivity and specificity of 67.6% (95% CI 49.1–82.6) and 92.9% (95% CI 66.1–99.8), respectively, in HIV-coinfected patients. In this setting, VL could be ruled-in with peripheral blood microscopy in a substantial number of VL suspects and may reduce the number of tissue aspirations performed, particularly in HIV-coinfected patients. More sensitive and logistically feasible methods than light microscopy are needed to detect Leishmania donovani parasites present in blood.

Introduction

Visceral leishmaniasis (VL) is a fatal infection, with an estimated 200,000–400,000 new cases and 20,000–40,000 deaths annually, second only to malaria in number of fatalities from parasitic diseases.1,2 The incidence and geographic range of VL are currently increasing in many areas, particularly in east Africa3,4 and overlap with several other tropical infectious diseases. Access to reliable diagnostic tests is one of the key factors limiting the care of VL patients and their communities. Clinical definitions of VL lack specificity, and the toxicity and high cost of therapy make confirmatory diagnostic tests essential. Unfortunately, both individual case management and VL control strategies are hampered by inadequate diagnostic tools. The widely available rK39 rapid diagnostic test (RDT) has been found to be insufficiently sensitive to rule out disease in east Africa, compared with other regions.5,6 This is especially true for HIV-coinfected individuals, in whom tests detecting host antibodies are frequently insensitive.7 Moreover, treatment of VL—with or without concomitant highly active antiretroviral therapy—is much more likely to fail than in non-HIV-coinfected patients, resulting in a relapse rate of 60% in the first year.8,9 These frequent relapses cannot be diagnosed using antibody detection tests because of long-term persistence of antibodies in those in whom they were initially present.10

Thus, microscopy or culture of splenic and bone marrow aspirates is the principal test used to confirm VL in east Africa and have good specificity. However, complex clinical infrastructure and expertise are required to perform them safely, thus limiting their use. Moreover, these procedures are associated with significant morbidity (e.g., bleeding and pain), are procedurally difficult in children, and are contraindicated in many situations inherent to the disease itself (e.g., thrombocytopenia, severe anemia). Lymph node aspirates are less invasive but also less sensitive11 and not applicable in all VL foci, as lymph node enlargement is not a common finding in most endemic regions outside Sudan.

Microscopic examination of peripheral blood has been reported to detect amastigote-stage parasites from Mediterranean patients with Leishmania infantum infection, in small, uncontrolled, non-blinded studies.1215 This may not apply to other contexts. Searching for a less invasive alternative to tissue aspirations, we sought to determine the diagnostic accuracy of peripheral blood microscopy for the diagnosis of VL using three blood processing techniques aimed at specimen concentration, among consecutive patients with clinically suspected VL from Leishmania donovani–endemic areas in Ethiopia. Smears of whole blood, buffy coat, and isolated peripheral blood mononuclear cells (PBMCs) were compared with a composite parasitological reference standard (microscopy of bone marrow and/or spleen aspiration).

Methods

Ethics statement.

This study was reviewed and approved by the Institutional Review Board of the Institute of Tropical Medicine, the ethics committee of the University Hospital in Antwerp, as well as the Institutional Review Board of the University of Gondar. All patients were included into the study after written informed consent was obtained. The management of VL diagnosed patients was done according to Ethiopian national guidelines. Patients found not to have VL were linked to the appropriate units in the hospital for further workup and management of their illness.

Study design.

A single-center study with a phase III diagnostic accuracy design (i.e., evaluation of test performance on a representative sample of prospectively recruited patients with clinically suspected VL)16 was performed to determine the accuracy of microscopic examination of peripheral blood for the diagnosis of VL. An independent study monitor assessed adherence to study methods 1 year into recruitment.

Study setting, population, and sampling strategy.

The study was carried out at University of Gondar Hospital (UoGH). The hospital is situated in the northwest part of Ethiopia close to the major VL-endemic sites along the Ethio–Sudanese border. There is a Leishmaniasis Research and Treatment Center (LRTC) in the hospital where several clinical trials on the treatment of leishmaniasis have been conducted. This center is supported by the Drugs for Neglected Diseases initiative (DNDi). Clinical assessment of patients followed by serology and splenic or bone marrow aspiration is part of routine care. The center has its own laboratory where Good Clinical Laboratory Practice trained and experienced technicians are working.

Young adult migrants moving from Leishmania-non-endemic highland areas to the Leishmania-endemic lowlands looking for work present with prolonged fever and symptoms of anemia. Patients from VL-endemic areas also present to the hospital seeking better care. Other diseases with a similar clinical picture as VL are also commonly seen in this geographic area. These include hyperreactive malarial splenomegaly, hepatosplenic schistosomiasis, and hematological malignancies.

Sample size and power.

Sample size calculations were based on the assumption that tests with even moderate sensitivity would be of clinical benefit for those patients in whom they yield a confirmed diagnosis and spare them more invasive procedures. This is especially important for primary health-care settings where performing the invasive procedures is often not possible.

Thus, using a precision-based calculation with an estimated sensitivity of 60% and a precision of 7% for each test under study, we aimed to recruit at least 189 VL cases. However, specificity must be very high for any proposed test regarded as confirmatory. Thus, with an estimated specificity of 95% and a precision of 5% for each test under study, we aimed to recruit at least 73 controls (i.e., people suspected of VL in whom the reference standard is negative). These numbers were increased to a target recruitment of 200 cases and 100 controls to allow for unexpected issues such as missing data or reversal of informed consent.

Subject-related interventions or procedures.

All consecutive VL suspects either referred or directly coming to the LRTC of UoGH were approached to consent and to participate in the study. Patients underwent routine clinical assessment and laboratory evaluation. This included clinical assessment of patients, followed by serological rK-39 testing, HIV testing, and splenic or bone marrow aspiration for microscopy. Tissue aspiration was performed for all clinical VL suspects, defined as fever and/or constitutional symptoms for more than 2 weeks and a palpable spleen in the absence of a confirmed alternative diagnosis in patients from the VL-endemic regions. In addition to tissue aspirates, smears of whole blood, buffy coat, and isolated PBMCs from each patient were Giemsa stained and examined. Two blinded readers independently assessed all slides and assessed parasite density according to the semilogarithmic World Health Organization (WHO) grading scale of 0 to 6.17 All fields were examined. Slides with discordant results were identified during data entry and were made to be re-read by both readers (blinded to their first reading) in the same room, and a consensus result was reached. Blinding to tissue aspirate results was maintained throughout. All slides were Giemsa stained at a pH of 7.2 to facilitate distinction between extracellular Leishmania amastigotes and Plasmodium sp. trophozoites.

Isolation of mononuclear cells.

Approximately 2.5 mL blood was carefully layered onto 2.5 mL Ficoll-Paque Premium density gradient media (1.077 mg/mL) (GE Healthcare, Uppsala, Sweden) in a sterile 15-mL conical centrifugation tube, without mixing, and centrifuged at 500 × g for 30 minutes at 18°C. The layer of PBMC was collected from the interface of the plasma and Ficoll into a new centrifugation tube for washing. The cells were washed in approximately three volumes of phosphate buffered saline solution. Suspension and centrifugation were done twice with high (400 × g) and lower (60–100 × g) speed for 10 minutes each time to increase the mononuclear cell recovery and facilitate the removal of the platelets, respectively. After removing the supernatant, the pellet was resuspended in 500 μL saline solution. Smears were prepared by placing 10 μL on each slide, air drying, and fixing with methanol.

Isolation of buffy coat with lysis of erythrocytes.

Another 2.5 mL blood sample was directly centrifuged at 900 × g for 10 minutes. After the upper plasma layer was gently aspirated and discarded, the buffy layer was gently aspirated. Lysis of erythrocytes was performed by adding 5 mL hypotonic saline (0.2%) for 20 seconds. An equal volume of hypertonic saline (1.6%) was then added to achieve an isotonic solution, followed by repeat centrifugation for 10 minutes at 900 × g. The supernatant was discarded, and thick smears were prepared by placing 10 μL resuspended pellet on each slide, air drying, and fixing with methanol.

Thin peripheral blood film preparation.

Thin smears of peripheral whole-blood were prepared using standard techniques, and fixed with methanol.

Data analysis.

Data were entered into a Microsoft Access database (Microsoft, Seattle, WA) using Epi Info™ 3.5.4 (Centers for Disease Control and Prevention, Atlanta, GA) with checks for completeness and consistency. Data were analyzed using STATA version 11.2 (StataCorp LP, College Station, TX). Sensitivity, specificity, positive predictive value and negative predictive value of each test were computed using the diagt command after the primary variables of interest were processed. Precision was assessed with 95% confidence intervals. Forest plots of diagnostic accuracy were generated using the metan command. The reference standard was defined as the demonstration of leishmania parasite in a bone marrow or splenic aspirate by microscopy. Baseline demographic and clinical variables were collected for all patients, expressed in terms of frequencies/percentages for categorical variables and median/interquartile range (IQR) for continuous variables, and compared between the VL and non-VL groups in a prespecified manner. A logistic regression model was developed post hoc to assess the factors associated with the visualization of amastigotes on PBMC smears. Finally, the agreement between the two readers was assessed using the kappa-statistic measure of agreement scale from 0 to 1.

Blinding to the results of tissue aspiration.

Great care was taken to ensure that all peripheral blood microscopists were blinded to the results of tissue aspiration. Each patient enrolled in the study was given a unique study number. All smears of peripheral blood-derived specimens were made using prelabeled slides. Preprinted labels contained only the “specimen type” and a unique randomly generated 5-digit number, created using the “INT(RAND()*100,000)” function in Microsoft Excel (Microsoft, Seattle, WA). Random slide numbers were traceable to individual study patients via a password-protected master file, accessible only to the study principal invesitigators (Ermias Diro and Cedric Yansouni).

Results

Recruitment and baseline characteristics.

A total of 301 consecutive patients presenting to the Leishmaniasis Research and Treatment Center of University of Gondar Hospital with signs and symptoms leading to clinical suspicion of VL were included in the study, of which 208 (69.3%) were parasitologically confirmed to have VL by microscopy of spleen or bone marrow aspirates (Table 1). Young adult males comprised the majority of the study population (median age of 25, IQR 21–30 years). Two hundred and thirty-seven patients (78.7%) were seasonal migrants to areas of high VL transmission intensity. Twenty-nine of them had VL in the past, based on self-reporting and/or past medical records in our center. Seventeen of these 29 patients (58.6%) were HIV coinfected.

Table 1

Baseline characteristics of study population

CharacteristicsVL confirmed (N = 208)Non-VL (N = 93)
Median age in years (IQR)25 (20–30)27 (23–32)
Sex: male207 (99.5)91 (97.8)
Original residency
 Seasonal migrants to VL-endemic area176 (84.6)61 (65.6)
 Previous VL in past19 (9.1)10 (10.8)
Symptoms
 Spontaneous bleeding26 (12.5)10 (10.8)
 Fever200 (96.2)88 (94.6)
 Jaundice14 (6.7)5 (5.4)
 Weight loss196 (94.2)86 (92.5)
Signs
 BMI (kg/m2), (median, IQR)16.2 (15.2–17.6)16.8 (15.9–18.2)
 Edema32 (15.4)2 (2.2)
 Spleen palpable202 (96.7)87 (93.5)
 Spleen size (cm below costal margin), (median, IQR)8 (5–11)6 (5–10)
 Lymphadenopathy4 (1.9)1 (1.1)
Comorbid conditions
 HIV infection34 (16.3)14 (15.1)
 Sepsis28 (13.6)5 (5.4)
 Malaria10 (4.8)7 (7.5)
 TB diseases6 (2.9)3 (3.2)
Laboratory
 Hemoglobin (g/dL), median (IQR)8.5 (7.1–9.9)9.6 (7.5–11.3)
 WBC Count (× 109 cells/L), median, (IQR)1.5 (1.1–2.0)2.5 (1.3–2.9)
 Platelet count (× 109/L), median (IQR)57 (38–88)88 (44–128)
 Pancytopenia146 (70.2)47 (30.8)
 CD4 count (cells/μL), median (IQR)63 (32–115.5) (N = 28)122 (69–133) (N = 9)
 Tissue aspirationN = 208N = 93
 Spleen145 (69.7)64 (69.6)
 Bone marrow63 (30.3)28 (30.4)

BMI = body mass index; HIV = human immunodeficiency virus; IQR = 25–75 interquartile range; TB = tuberculosis; VL = visceral leishmaniasis; WBC = white blood cell. Numbers in brackets are percentages unless otherwise noted.

Clinical characteristics were comparable between the VL and control groups, with frequent fever, weight loss, and splenomegaly (Table 1). Most patients were malnourished with a median BMI of 16.2 kg/m2 and 16.8 kg/m2 in the VL and control groups, respectively. Pancytopenia—defined as the combination of severe anemia (hemoglobin < 10 g/dL or hematocrit <30%), leukopenia (white blood cell count < 4 × 109/L), and thrombocytopenia (platelet count < 150 × 109/L)—was more common in VL patients than controls (70.2% versus 30.8%, respectively). HIV prevalence was 48/301 (15.9%) overall and 34/208 (16.3%) among the VL-confirmed patients. Besides HIV, bacterial sepsis in 33/301 (10.9%) and malaria in 17/301 (5.6%) were the most commonly detected infectious comorbidities. Nine of 301 (2.9%) patients had tuberculosis disease. Around 30% of tissue aspirations were from bone marrow, because of contraindications to spleen puncture, such as spontaneous bleeding, severe thrombocytopenia (platelet count less than 40 × 109/L), severe anemia (hemoglobin less than 3.0 g/dL), or inability lie still during the procedure for any reason.

Diagnostic accuracy of peripheral blood microscopy according to parasite concentration method.

Results according to parasite concentration method are shown in Figure 1. Compared with microscopy of tissue aspirates, smears of isolated PBMCs were significantly more likely to yield amastigotes than the other methods, with accurate identification of 70/208 confirmed VL patients (sensitivity 33.7% [95% CI 27.3–40.5]), and 88/93 clinical suspects found not to have VL (specificity 95.7% [95% CI 89.4–98.8]). Buffy coat correctly identified 40/205 tissue aspirate microscopy-positive cases (sensitivity 19.5% [95% CI 14.3–25.6]) and 91/92 clinical suspects without VL (specificity 98.9% [95% CI 94.1–100]). Finally, thin smear preparations of peripheral whole blood revealed amastigotes in only 1/65 people with VL (sensitivity 1.5% [95% CI 0.0–8.3]), and 0/14 clinical suspects without VL (specificity 100% [95% CI 76.8–100]). Use of this technique was discontinued after 79 people were recruited, owing to its low sensitivity.

Figure 1.
Figure 1.

Diagnostic accuracy of peripheral blood microscopy compared with invasive tissue aspirate microscopy, according to (A) parasite concentration method (B) and human immunodeficiency virus (HIV) status. FN = false negative; FP = false positive; HIV-neg = HIV-seronegative patients; HIV-pos = HIV-seropositive patients; PBMC = peripheral blood mononuclear cell; TN = true negative; TP = true positive. Numbers in brackets and bars represent 95% confidence interval (CI). Combined PBMC/buffy denotes the cumulative result from smears of both PBMC and the buffy coat of a given patient.

Citation: The American Society of Tropical Medicine and Hygiene 96, 1; 10.4269/ajtmh.16-0362

Twenty-six of the 138 VL cases (18.8%) missed by PBMC concentration were positive by buffy-coat smear. Conversely, 54/165 VL cases (32.7%) missed by buffy coat were positive by PBMC microscopy (Figure 2). By combing the results from both techniques, 96/208 cases were detected (sensitivity 46.2% [95% CI 39.2–53.2]), and 89/93 true negatives were identified (specificity 95.7% [95% CI 89.4–98.8]) (Figure 1A). Figure 3 shows the appearance of L. donovani amastigotes in the spleen, PBMC, and buffy coat of VL patients.

Figure 2.
Figure 2.

Overlap of results from microscopic examination of tissue aspirates, PBMC isolation, and buffy coat. Note that buffy-coat smears from 4/301 patients during the study were either lost or not examined, yielding a total of 297 patients depicted. Percentages in the circles are of the total population. PBMC = peripheral blood mononuclear cells.

Citation: The American Society of Tropical Medicine and Hygiene 96, 1; 10.4269/ajtmh.16-0362

Figure 3.
Figure 3.

Leishmania donovani amastigotes in the (A and B) spleen, (C) PBMC, and (D) buffy coat of visceral leishmaniasis patients, using Giemsa stain. Arrowheads indicate amastigotes with visible nucleus and kinetoplast.

Citation: The American Society of Tropical Medicine and Hygiene 96, 1; 10.4269/ajtmh.16-0362

In a subgroup analysis, the sensitivity and specificity of PBMC smears in HIV-coinfected patients were 55.9% (95% CI 37.9–72.8) and 92.9% (95% CI 66.1–99.8), whereas it was 29.3% (95% CI 22.7–36.7) and 94.9% (95% CI 89.3–99.2), respectively, in non-HIV-coinfected patients (Figure 1B). The sensitivity was significantly higher among HIV-coinfected groups, two-sample test of proportions, P = 0.0097. Combining the results from smears of both PBMC and buffy coat yielded a sensitivity and specificity of 67.6% (95% CI 49.1–82.6) and 92.9% (95% CI 66.1–99.8), respectively, in HIV-coinfected patients and a sensitivity and specificity of 41.7% (95% CI 34.3–49.4) and 96.2% (95% CI 89.3–99.2), respectively, in non-HIV-coinfected patients.

Associations between peripheral blood positivity, tissue aspirate grade, and HIV status.

HIV-coinfected patients tended to have a higher tissue parasite grade on the WHO semilogarithmic scale of 0 to 6 (Pearson χ2 P < 0.0001). In a logistic regression model, increasing tissue-aspirate grade in turn was associated with higher probability of seeing amastigotes on PBMC smears, independent of HIV status (odds ratio [OR] 2.17 [95% CI 1.77–2.66]).

Interobserver agreement.

There was moderate agreement in overall buffy-coat reading (kappa = 0.5119) and fair agreement among the aspirate negative groups (kappa = 0.3109). The degree of agreement in the overall Ficoll concentrate (PBMC isolates) was moderate (kappa = 0.4517), and slight (kappa = 0.1410) when aspiration is negative. The mean duration spent on single slide reading was 40 (standard deviation [SD] 7.4) and 47 (SD 7.3) minutes by the first and second readers, respectively. Reading times in the study were felt to be compatible with those in clinical practice, but these are not systematically recorded in routine work. Table 2 summarizes results from individual readers and the final “consensus” result.

Table 2

Sensitivity and specificity results from individual readers and the final “consensus” result from rereading discordant slides by both readers together

 PBMC isolationBuffy coatWhole blood thin smear
Sensitivity (%)Reader 133.218.91.54
Reader 231.922.90
Consensus33.719.5n/a
Specificity (%)Reader 184.896.7100
Reader 294.696.7100
Consensus95.798.9n/a

n/a = not applicable; PBMC = peripheral blood mononuclear cell. Readers remained blinded to the tissue aspirate results during “consensus” reading.

Discussion

Microscopy of the buffy-coat portion of whole blood is one of the parasitological diagnostic methods for VL recommended by WHO.17 This is based on small uncontrolled studies from the Mediterranean region with L. infantum infection, in which sensitivities of up to 56–68% were reported.12,14,15,18 To our knowledge, this study is the first adequately powered evaluation of parasite concentration techniques for the diagnosis of VL comprising only prospectively recruited VL suspects, with and without VL, subjected to the same reference standard, and in which readers were meticulously blinded to true disease status.

Patients in our study were all at advanced stages of illness and represent the severe end of the clinical spectrum of VL. In this group, PBMC isolation was significantly superior to buffy-coat preparations, whereas thin smears of peripheral blood were found to be of no diagnostic utility for leishmaniasis. The sensitivity of all methods was far lower than that in previously published reports.1215 On the basis of our experience, this likely reflects the fact that previous accounts were only among patients known to have VL, in which objects of uncertain morphology might have been considered positive. In reality, therapeutic decisions for VL involve toxic or costly treatments. This means that test specificity must be very high (e.g., by requiring the visualization of several objects demonstrating both a nucleus and typical kinetoplast for a slide to be considered positive), and this stringency requirement limits the sensitivity achievable by morphological assessment.

Another challenge to diagnostic accuracy was the low number of parasites seen on each slide, with all blood-derived positive slides in the study having a grade of 1+ (i.e., 1–10 parasites/1,000 high-power fields) or 2+ (i.e., 1–10 parasites/100 high-power fields) on the WHO scale.17 In keeping with previous reports of quality-assurance programs for VL,19 parasitemia grade correlates with interobserver agreement. Thus, the low parasitemia in these methods likely explains many of the discordant results we observed between readers.

The reference standard used in this study was Giemsa-stained microcopy of tissue aspirations. Although two-third of the study patients had splenic aspiration, the rest underwent bone marrow aspiration, which has lower diagnostic yield. This imperfect reference might also have affected the results.

Overlap between tissue aspirates, PBMC isolation, and buffy coat.

We hypothesized at study outset that PBMC isolation would yield a higher parasite density than buffy-coat preparations, based on the fact that amastigotes in vivo are chiefly intracellular parasites within macrophages/monocytes. Although this work supports our hypothesis, we also found that a substantial number of parasites were visualized by both readers in smears of buffy coat but not of PBMC (Figure 2). These findings suggest that the density of parasitized PBMC and/or extracellular parasites is variable, and that there is a significant diagnostic yield to examining preparations of both PBMC and buffy coat.

HIV-coinfected subgroup.

A large difference in the sensitivity of PBMC preparations was observed when stratifying by HIV status (Figure 1B). This may be attributable to impaired immunity and correlates with the higher tissue–parasite burden we observed in HIV-coinfected patients. Higher parasite load may indicate prolonged duration of illness, severity of clinical disease, and higher mortality. In that case, PBMC isolation might help to accurately diagnose VL while avoiding invasive tissue aspiration in critically sick patients. Although the association of high parasitemia with clinical disease severity is not clearly established, high parasite burden in tissues is predictive of VL treatment failure among HIV-coinfected patients.20

Impact of the study design on diagnostic accuracy results.

Several features of this study were aimed at minimizing sources of bias. First, all patients presenting with features compatible with VL were recruited to the study over a period of 2 years. Diagnostic methods were applied to all diseases suspects, resulting in the most representative “negative control” patients possible. Investigators went to great lengths to ensure that study microscopists remained blinded to tissue aspiration results while reading smears of PBMC, buffy coat, and whole blood. Moreover, fixed and stained slides were read over a period of several months instead of real time, further reducing the probability of associating a given slide with others from a particular patient. An independent study monitor assessed adherence to study methods 1 year into recruitment, and found no major nonconformities with standard operating procedures and safeguards. As such, we do not believe that prior knowledge of true disease status was a significant source of bias in this work. However, the large number of slides generated (over 1,200 in total) and considerable time spent on reading each one (mean 43.5 minutes) forced us to apply a third test only for slides with discordant interpretations from the two microscopists. The resulting specificities generated in the final consensus results are slightly higher than those obtained by either individual reader (Table 2). This effect on specificity is a known caveat to the use of “discrepant analysis.”2123 Nonetheless, we believe that these results are valid because the comparison here is not between a reference standard and comparator, but between two readers of comparable skill reflecting real-world practice in laboratories, where technologists regularly consult each other when facing a difficult unknown specimen.

Overall conclusions, feasibility, and future directions.

This study establishes with a high degree of confidence that a substantial number of patients with VL can be parasitologically diagnosed on the basis of peripheral blood microscopy, and that these patients may be spared more invasive diagnostic procedures. We also demonstrate that microscopic examination of PBMCs is more sensitive than examination of the buffy coat, and that the combination of PBMC and buffy-coat preparations increases the number of VL-infected people detected by less invasive techniques. Avoiding invasive procedures may also benefit health-care workers by reducing the risk of occupational injuries.

Unfortunately, each blood specimen requires approximately 2.5 hours of processing before microscopy, compared with a few minutes of processing for invasive tissue aspirates. This fact, along with generally modest sensitivity and the high expertise required, limit the techniques described in this study to be applicable only at expert referral centers, and be considered only in carefully selected patients.

The demonstrable presence of parasites in peripheral blood highlights the need for more sensitive detection methods that can be used more quickly and easily than light microscopy. Examples could include fluorescence-based microscopy, antigen detection, molecular assays, and—for certain indications—blood-based parasite culture of isolated PBMC specimens.

ACKNOWLEDGMENTS

We wish to thank the patients recruited to this study, as well as staff at the Leishmaniasis Research and Treatment Center and School of Biomedical and Laboratory Sciences of the University of Gondar for their technical assistance and facilitation of this work.

  • 1.

    Alvar J, Velez ID, Bern C, Herrero M, Desjeux P, Cano J, Jannin J, den Boer M, 2012. Leishmaniasis worldwide and global estimates of its incidence. PLoS One 7: e35671.

    • Search Google Scholar
    • Export Citation
  • 2.

    van Griensven J, Diro E, 2012. Visceral leishmaniasis. Infect Dis Clin North Am 26: 309322.

  • 3.

    Alvar J, Bashaye S, Argaw D, Cruz I, Aparicio P, Kassa A, Orfanos G, Parreno F, Babaniyi O, Gudeta N, Canavate C, Bern C, 2007. Kala-azar outbreak in Libo Kemkem, Ethiopia: epidemiologic and parasitologic assessment. Am J Trop Med Hyg 77: 275282.

    • Search Google Scholar
    • Export Citation
  • 4.

    Bern C, Maguire JH, Alvar J, 2008. Complexities of assessing the disease burden attributable to leishmaniasis. PLoS Negl Trop Dis 2: e313.

  • 5.

    Boelaert M, Verdonck K, Menten J, Sunyoto T, van Griensven J, Chappuis F, Rijal S, 2014. Rapid tests for the diagnosis of visceral leishmaniasis in patients with suspected disease. Cochrane Database Syst Rev 6: CD009135.

    • Search Google Scholar
    • Export Citation
  • 6.

    Cunningham J, Hasker E, Das P, El Safi S, Goto H, Mondal D, Mbuchi M, Mukhtar M, Rabello A, Rijal S, Sundar S, Wasunna M, Adams E, Menten J, Peeling R, Boelaert M, 2012. A global comparative evaluation of commercial immunochromatographic rapid diagnostic tests for visceral leishmaniasis. Clin Infect Dis 55: 13121319.

    • Search Google Scholar
    • Export Citation
  • 7.

    Medrano FJ, Canavate C, Leal M, Rey C, Lissen E, Alvar J, 1998. The role of serology in the diagnosis and prognosis of visceral leishmaniasis in patients coinfected with human immunodeficiency virus type-1. Am J Trop Med Hyg 59: 155162.

    • Search Google Scholar
    • Export Citation
  • 8.

    ter Horst R, Collin SM, Ritmeijer K, Bogale A, Davidson RN, 2008. Concordant HIV infection and visceral leishmaniasis in Ethiopia: the influence of antiretroviral treatment and other factors on outcome. Clin Infect Dis 46: 17021709.

    • Search Google Scholar
    • Export Citation
  • 9.

    Diro E, Lynen L, Ritmeijer K, Boelaert M, Hailu A, van Griensven J, 2014. Visceral Leishmaniasis and HIV coinfection in east Africa. PLoS Negl Trop Dis 8: e2869.

    • Search Google Scholar
    • Export Citation
  • 10.

    Gidwani K, Picado A, Ostyn B, Singh SP, Kumar R, Khanal B, Lejon V, Chappuis F, Boelaert M, Sundar S, 2011. Persistence of Leishmania donovani antibodies in past visceral leishmaniasis cases in India. Clin Vaccine Immunol 18: 346348.

    • Search Google Scholar
    • Export Citation
  • 11.

    Boelaert M, el Safi S, Goetghebeur E, Gomes-Pereira S, Le Ray D, Van der Stuyft P, 1999. Latent class analysis permits unbiased estimates of the validity of DAT for the diagnosis of visceral leishmaniasis. Trop Med Int Health 4: 395401.

    • Search Google Scholar
    • Export Citation
  • 12.

    Chemli J, Abroug M, Fathallah A, Abroug S, Ben SM, Harbi A, 2006. Contribution of leukoconcentration in the diagnosis of Kala-azar in Tunisia [in French]. Med Mal Infect 36: 390392.

    • Search Google Scholar
    • Export Citation
  • 13.

    Deniau M, Canavate C, Faraut-Gambarelli F, Marty P, 2003. The biological diagnosis of leishmaniasis in HIV-infected patients. Ann Trop Med Parasitol 97 (Suppl 1): 115133.

    • Search Google Scholar
    • Export Citation
  • 14.

    Izri MA, Robineau M, Petithory JC, Rousset JJ, 1993. Visceral leishmaniasis. Parasitological diagnosis by leukocyte concentration [in French]. Presse Med 22: 1010.

    • Search Google Scholar
    • Export Citation
  • 15.

    Izri MA, Deniau M, Briere C, Rivollet D, Petithory JC, Houin R, Rousset JJ, 1996. Leishmaniasis in AIDS patients: results of leukocytoconcentration, a fast biological method of diagnosis. Bull World Health Organ 74: 9193.

    • Search Google Scholar
    • Export Citation
  • 16.

    Boelaert M, Bhattacharya S, Chappuis F, el Safi S, Hailu A, Mondal D, Rijal S, Sundar S, Wasunna M, Peeling RW, 2007. Evaluation of rapid diagnostic tests: visceral leishmaniasis. Nat Rev Microbiol 5: S30S39.

    • Search Google Scholar
    • Export Citation
  • 17.

    World Health Organization, 2010. Control of the Leishmaniases. Report of a meeting of the WHO Expert Committee on the Control of Leishmaniases, Geneva, March 22–26, 2010. Geneva, Switzerland: WHO Technical Report Series 949, World Health Organization.

    • Search Google Scholar
    • Export Citation
  • 18.

    Petithory JC, Ardoin F, Ash LR, Vandemeulebroucke E, Galeazzi G, Dufour M, Paugam A, 1997. Microscopic diagnosis of blood parasites following a cytoconcentration technique. Am J Trop Med Hyg 57: 637642.

    • Search Google Scholar
    • Export Citation
  • 19.

    Boelaert M, Rijal S, Regmi S, Singh R, Karki B, Jacquet D, Chappuis F, Campino L, Desjeux P, Le Ray D, Koirala S, Van der Stuyft P, 2004. A comparative study of the effectiveness of diagnostic tests for visceral leishmaniasis. Am J Trop Med Hyg 70: 7277.

    • Search Google Scholar
    • Export Citation
  • 20.

    Diro E, Lynen L, Mohammed R, Boelaert M, Hailu A, van Griensven J, 2014. High parasitological failure rate of visceral leishmaniasis to sodium stibogluconate among HIV co-infected adults in Ethiopia. PLoS Negl Trop Dis 8: e2875.

    • Search Google Scholar
    • Export Citation
  • 21.

    Green TA, Black CA, Johnson RE, 2001. In defense of discrepant analysis. J Clin Epidemiol 54: 210215.

  • 22.

    Lipman HB, Astles JR, 1998. Quantifying the bias associated with use of discrepant analysis. Clin Chem 44: 108115.

  • 23.

    Schachter J, 2001. In defense of discrepant analysis. J Clin Epidemiol 54: 211212.

Author Notes

* Address correspondence to Cédric Yansouni, Division of Infectious Diseases, Department of Microbiology, J.D. MacLean Centre for Tropical Diseases at McGill University, McGill University Health Centre, 1001 Décarie Bouelvard, Room EM3.3242, Montreal, QC, H4A 3J1, Canada. E-mail: cedric.yansouni@mcgill.ca† These authors contributed equally to this work.

Financial support: This study was supported by a grant from AIDS Mecenaat at Institute of Tropical Medicine (ITM). Ermias Diro has received a PhD scholarship granted from the Belgian Directorate General for Development Cooperation under the ITM-DGDC framework agreement FA-III and from the European Union Seventh Framework Programme (FP7/2007-2013) under grant agreement no. 305178 via AfriCoLeish project. Cedric Yansouni holds a “Chercheur-boursier clinicien” career award from the Fonds de recherche du Québec—Santé (FRQS).

Authors' addresses: Ermias Diro, University of Gondar, Gondar, Ethiopia, and Institute of Tropical Medicine, Antwerp, Belgium, E-mail: ermi_diro@yahoo.com. Cedric Yansouni, J.D. MacLean Centre for Tropical Diseases at McGill University, Montreal, Canada, E-mail: cedric.yansouni@mcgill.ca. Yegnasew Takele and Bewketu Mengesha, University of Gondar, Gondar, Ethiopia, E-mails: yegnasew77@yahoo.com and bewketmd@yahoo.com. Lutgarde Lynen, Johan van Griensven, Marleen Boelaert, and Philippe Büscher, Institute of Tropical Medicine, Antwerp, Belgium, E-mails: llynen@itg.be, jvangriensven@itg.be, mboelaert@itg.be, and pbuscher@itg.be. Asrat Hailu, School of Medicine, Addis Ababa University College of Health Sciences, Addis Ababa, Ethiopia, E-mail: hailu_a2004@yahoo.com.

Save