Ultrastructure of the Liver and the Spleen in Fatal Malaria

We performed a retrospective study of 25 patients who died of severe falciparum malaria in Thailand and Vietnam using electron microscopy. The aims of the study were: to determine if there was any significant association between parasitized red blood cells (PRBC) sequestered in liver and spleen and particular pre-mortem clinical complications, and to compare the degree of parasite load between the liver and spleen within the same patients. PRBC sequestrations in each organ were compared with the pre-mortem parasitemia, to calculate the sequestration index (S.I.). The S.I. showed that the degree of PRBC sequestration in the spleen was higher than the liver (S.I. median = 3.13, 0.87, respectively) (p < 0.05). The results of quantitative ultrastructural study showed a significantly high parasite load in the liver of patients with jaundice, hepatomegaly and liver enzyme elevation (p < 0.05). We found a significant correlation between PRBC sequestration in the liver and a high serum bilirubin level, a high aspartate aminotransferase (AST) level and an increase in the size of the liver (Spearman’s correlation coefficient = 0.688, 0.572, 0.736, respectively). Furthermore, a higher parasite load was found in the liver of patients with acute renal failure (ARF) compared to patients without ARF (p<0.05). These findings suggest that PRBC sequestration in the liver is quantitatively associated with pre-mortem hepatic dysfunction and renal impairment. There was no significant difference between splenomegaly and PRBC sequestration. The size of a palpable spleen was not correlated with parasite load in the spleen. When ultrastructural features were compared between the two reticuloendothelial organs, we found that the spleen had more PRBC and phagocytes than the liver. The spleen of non-cerebral malaria (NCM) patients had more phagocytes than cerebral malaria (CM) patients. This observation reveals that the spleen plays a major role in malaria parasite clearance, and is associated with host defence mechanisms against malaria. questration of PRBC in specific organs has been proposed as an important pathophysiological factor in clinical patterns of severe malaria. CM accounts for many of the deaths in severe malaria infected patients. Most previous studies have tended to concentrate on the pathophysiology of changes in the brain (Macpherson et al, 1985; Pongponratn et al, 1991.) The degree of sequestration of parasitized red blood cells (PRBC) in cerebral microvessels has been found to be associated with pre-mortem coma (Pongponratn et al, 2003). There has been no reported study of the ultrastructure focusing on the association between the sequestration of PRBC in other organs and the particular clinical SOUTHEAST ASIAN J TROP MED PUBLIC HEALTH 1360 Vol 36 No. 6 November 2005 complications. Several studies have reported abnormal liver and spleen function during malaria, indicated by hepatomegaly (Sowunmi et al, 2001), splenomegaly (Looareesuwan et al, 1987) and liver enzyme elevation (Premaratna et al, 2001; Kochar et al, 2003). Jaundice has often been reported in patients with cerebral malaria, hyperparasitemia, acute renal failure and shock (Wilairatana et al, 1994). The pathology that causes hepatic dysfunction and the role of the spleen in Plasmodium falciparum malaria in humans is still unclear. The main objectives of this study were to demonstrate PRBC sequestration in reticuloendothelial organs, including the liver and spleen, and to determine whether premortem clinical complications, such as jaundice and hepatosplenomegaly were associated with PRBC sequestration in these organs. The other objectives were to correlate the clinical complications with pathological f indings in fatal falciparum malaria and to examine the interactions between PRBC and host cells, including endothelial cells, leukocytes and platelets. MATERIALS AND METHODS Patients Specimens were collected from patients who died of severe falciparum malaria in Thailand and Vietnam from 1987 to 2001. A full autopsy was performed on these patients with routine pathological examinations. When autopsy was not possible, percutaneous needle biopsy was performed on some cases. Informed consent for autopsy and tissue necropsy were obtained from the patients or their accompanying relatives. Ultrastructural studies were done without the knowledge of clinical information until the end of the statistical analysis. Ultrastructural studies The tissues were cut and fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer at pH 7.4. Fixed tissues were subsequently washed three times with 0.1 M phosphate buffer at pH 7.4, post fixed with 1% osmium tetroxide in 0.1 M phosphate buffer at pH 7.4, at 4oC for 1 hour, dehydrated in a graded ethanol series, infiltrated with propylene oxide, and embedded in Epon 812 or spurr’s epoxy resin. Ultrathin sections were cut with glass knives, mounted on copper grids (200-mesh square), stained with 2% uranyl acetate and Reynold’s lead citrate prior to examine in a transmission electron microscope (Hitachi H-7000 model). Qualitative and quantitative ultrastructural analysis was performed on ten grid squares of the copper grids depending on the size of the tissues. Qualitative ultrastructural analysis The presence or absence of sequestered PRBC, leukocytes, fibrin thrombi, platelets and endothelial cells changes was recorded. Electron micrographs were taken of relevant areas. Quantitative ultrastructural analysis The number of NRBC and PRBC, which were identified as knob positive PRBC(K+PRBC), knob negative PRBC (K-PRBC) and schizonts (defined as multisegmented parasites with more than three nuclei per PRBC) were counted. Only macrophages, white blood cells or endothelial cells that showed phagocytosed NRBC, PRBC or malarial pigments were counted as phagocytes. All counting parameters were recorded in terms of the number counted per grid square. The parasite load in each organ were calculated and expressed as a sequestration index (S.I.). The S.I. has been previously described by Silamut et al (1999) as a way to estimate the difference between PRBC accumulation in a tissue vascular bed, compared to the peripheral blood parasitemia. The S.I. was calculated from the ratio of the percentage of PRBC in a tissue vascular bed to the percentage of PRBC in the peripheral blood. Statistical methods Data were analyzed by SPSS for Windows version 11.0. To compare medians between the different clinical groups, we used the MannWhitney U test. For comparison of the percentages of cases with and without PRBC between different clinical groups. The chi-square test or Fisher’s exact test were used. To compare the parasite load between the liver and spleen within the same patients, the Wilcoxon signed rank test was used. We used Spearman’s correlation coefficient to explain the correlation between PRBC sequestration in the liver and the level of serum bilirubin or liver enzymes. ULTRASTRUCTURE OF THE LIVER AND THE SPLEEN IN FATAL MALARIA Vol 36 No. 6 November 2005 1361 RESULTS Summary of clinical data The clinical summaries of the malaria cases are shown in Table 1. There were 25 patients included in this study. Of these, 19 liver tissues and 14 spleen tissues were collected. Only 8 patients had tissue samples collected from both organs were compared. Qualitative ultrastructural analysis Degenerative post-mortem changes were seen. However despite these changes, the presence of obvious pathological processes was observed. Ultrastructural features of the spleen showed congestion of splenic sinusoids and the splenic artery with normal and PRBC (Fig 1A). Splenic pitting of PRBC was commonly seen (Fig 1B and 1C). Ultrastructural features of the liver showed less congestion compared to the spleen. Hepatic sinusoids were filled with NRBC, PRBC, schizonts and phagocytes. In these patients, all parasite stages included K-PRBC, K+PRBC and schizonts. Occasionally gametocytes were found; many of them had an irregular shape, characterized by distortion and elongation of cytoplasmic processes (Fig 1D). Platelets were occasionally found. It was rare to find fibrin deposition (Fig 1E). Phagocytosis of RBC, PRBC and malarial pigment was always found in both reticuloendothelial organs and more predominantly was associated with mononuclear cells (MNC) (Fig 1F). Reactive Kupffer’s cells containing PRBC, RBC or malarial pigment were commonly seen (Fig 2A). There was clear evidence of PRBC adhering to sinusoidal endothelial cells observed in the liver and spleen of this study (Fig 2B). Minimal changes in the endothelial cells were seen. Quantitative ultrastructural analysis The spleen. To determine whether the pathological features of the spleen were associated with the complications of falciparum malaria, the counting parameters of the spleen were compared among the different clinical groups. Comparative quantitation of the counting parameters per grid square of the spleen and in the different clinical groups is shown in Table 2. Statistical analysis showed no differences in the number of grid squares counted between the patient groups. There was no significant difference in most of the counting parameters examined of the spleen in the clinical groups except for the number of phagocytes, which was significantly higher in the NCM and jaundice patients compared to the CM and patients with no jaundice (p=0.024, p=0.030, respectively). The liver. Comparative quantitation of the counting parameters examined in the liver among the different clinical groups is shown in Table 3. Statistical analysis showed no differences in the number of grid squares counted in the patient groups. There were no significant differences in most of the parameters examined in the livers of CM patients compared to NCM patients. Comparing patients with and without jaundice, the following parameters were significantly higher in patients with jaundice compared to patients without jaundice: the percentage of PRBC (p=0.028), S.I. of PRBC (p=0.011) and S.I. of KPRBC (p=0.008). The chi-square test and Fisher’s exact test were used to compare percentages, and no significant associations were found between patients with jaundice and the percentage of cases with PRBC and K-PRBC (p=0.005 and p=0.007, respectively). The clinical and laboratory findings of malaria patients are given in Table 4. Blood chemistry showed an elevation in serum bilirubin (mean 29.28 mg/dl; range 4.22-175 mg/dl) in patients with jaundice. There was a significant correlation between the S.I. of PRBC and the level of serum bilirubin, including the total and direct bilirubin (Spearman’s correlation coefficient=0.688, 0.617; p=0.002, 0.006, respectively; Table 3). Significantly higher percentages of PRBC and K-PRBC were found in patients with liver enzyme elevation compared to those without liver enzyme elevation (p=0.033 and p=0.042, respectively; Table 3). There was a significant correlation between the percentage of PRBC in the liver and the level of aspartate transaminase (AST) (mean = 88.23 U/ml; range 22-291 U/ml; Spearman’s correlation coefficient = 0.572; p=0.026; Table 3). As with others (Nacher et al, 2001b), we considered a palpable liver as a reflection of hepatomegaly. Because a palpable liver is not the best measure of hepatomegaly, we defined SOUTHEAST ASIAN J TROP MED PUBLIC HEALTH 1362 Vol 36 No. 6 November 2005 1 25 M A 13 .5 96 7, 12 0 81 5, 50 0 (2 9. 51 ) C M 45 .6 1 28 .1 4 A R F, A R D S , H ep , J , S , H yp og ly 2 16 M A 1. 5 N A N A C M N A 2. 48 A R F, P ul m E d, C 3 61 M A 29 6. 5 3, 69 2, 64 0 1, 19 5, 33 5 (3 0. 7) C M 6. 49 N A A R F, P ul m E d, H yp er P 4 22 M A 17 18 9, 02 8 8, 54 0 (0 .2 ) C M 0 N A A R D S 5 55 F A 77 57 ,2 50 16 0 (0 .1 ) C M 1. 11 N A A R F, A R D S , H ep , J , H yp og ly 6 38 F A 80 1, 27 1, 16 0 80 (0 .3 ) C M 0 1. 39 N on e 7 40 M Q 38 58 ,6 96 45 5, 40 0 (1 3. 95 ) C M 9. 61 N A N on e 8 19 F Q 53 21 5, 17 6 29 ,6 01 (1 .2 ) N C M 4. 5 N A H ep , J , A n, D IC , s ep si s 9 20 M Q 3 1, 36 0, 12 8 1, 36 0, 12 8 (4 4. 8) N C M 13 .5 N A J, A n, H yp er P 10 45 M Q 53 .5 16 0, 45 4 0 C M 0. 87 2. 79 A R F, J , S 11 67 M Q 24 3, 03 0 12 ,4 34 (0 .3 ) C M 0. 69 N A N on e 12 26 F Q 8 44 4, 62 4 44 4, 62 4 (1 7. 5) C M 6. 27 10 .0 8 A R F, J , S 13 19 M Q 16 72 ,8 48 72 ,8 48 (2 .8 7) C M 17 .8 30 .6 4 A R F, J , A n, A ci do si s 14 27 M Q 81 1, 24 4, 94 7 6, 42 6 (0 .2 ) C M 1. 44 7. 17 J 15 57 M Q 57 17 9, 50 3 16 ,1 92 (0 .3 4) C M 8. 58 N A H ep , J , S 16 26 M Q 31 66 ,8 82 56 4, 57 2 (1 8. 73 ) C M 8. 11 N A A R F, J , S , H yp er P 17 63 M A 23 84 ,4 03 56 0 (0 .0 15 ) C M N A 24 .3 7 H ep , J , S ,C 18 34 M Q 4 80 5, 91 2 80 5, 91 2 (2 2. 3) N C M N A 45 .5 J, C 19 28 M A 4 54 6, 36 0 95 8, 83 0 (3 4. 9) C M N A 22 .8 2 H ep , S , H yp er P 20 28 M Q 44 61 9, 20 8 12 8, 86 5 (5 .4 ) N C M N A 16 .8 8 A n, S , C 21 28 M Q 95 .7 5 45 ,3 42 0 C M 20 .9 5 2. 18 H ep , J , A n, H yp og ly 22 34 M Q 6. 6 15 0, 72 0 53 ,8 82 (3 .3 ) C M 33 .2 2 23 .2 9 A R F, J , S , A n 23 22 F Q 27 .3 80 2, 58 4 56 ,5 20 (1 .8 ) N C M N A 20 .6 5 P ul m E d , H ep , J , S ,H yp er P 24 56 F A 4. 66 21 ,1 01 11 7, 09 6 (4 .7 ) N C M 0 N A N on e 25 28 M A 18 .5 62 9, 25 6 14 8, 83 6 (7 .9 ) N C M 3. 98 N A P ul m E d , J , A n, H yp og ly A = a rt em is in in d er iv at iv es , A n = a ne m ia , A R D S = a cu te r e sp ira to ry d is tr es s sy nd ro m e, A R F = a cu te r e na l f ai lu re , C = c om a, C M = c er eb ra l m al ar ia , H ep = h ep at om ega ly , H yp er P = h yp er p ar as ite m ia , H yp og ly = h yp og ly ce m ia , J = ja un d ic e, N A = n ot a va ila b le , P ul m E d = p ul m on ar y ed em a, N C M = n on -c er eb ra l m al ar ia , Q = q ui ni ne , S = s ho ck Ta b le 1 C lin ic al d et ai ls o f th e m al ar ia c as es (n = 2 5 ).