Although survival has improved in SCD, patients still die prematurely, especially males; and in this retrospective study, the overall proportion of death among males was double that seen in females. Similarly, the median age in males was lower (26.5 v/s 28 years), as compared to the females, but this difference was not statistically significant. Comparable findings were also reported in other studies from the region and beyond [14,15,18–20].
Karacaoglu PK et al [19], from Turkey, demonstrated that there were more deaths among males, than females, and the mean age of death in their female population was 40.1 years. Although patients are still dying young, data from a French SCD cohort in children under the age of 5 years showed that there is an improvement in survival attributed to the establishment of what is known as complete comprehensive patient care for SCD [20]. Al-Suliman et al,[21] in a study on adults from the eastern province in Saudi Arabia reported a higher prevalence of males 55.8% versus females 44.2% female, with the mean age of the male patients being 30 ± 14 years (range, 16–67 years) and of the female patients 27 ± 13 years (range, 14–67 years). The first Brazilian single institutional study in 2017 showed that the mortality rate was 18.87% among adult SCD patients [22]. Further, combined data between 1997–2017, from Brazil showed that mortality was more in the males (50.4%), and patients aged between 25 and 34 years had a higher incidence of deaths [23]. In another study from Brazil between 2000 to 2018, during the entire period the mean ages at death is significantly lower for males, than for females, being 29.4 (± 19.6) versus 33.3 (± 20.3) years respectively [24]. In a study from the UK, the CCU mortality for SCD patients was 19.6% between 2000 to 2007, with the mean ages for males and females respectively being 32.6 and 34.25 years [25]. Further, in a study from the USA with data spanning over 27 years between 1979 to 2005, the mean age at death was significantly different for males (33.4 years) than for females (36.9 years) [26]. Nonetheless, in recent years, patients are reported to live into their eighth decade as reported by Ballas SK,[27] and indeed in this study cohort, we had an SCD patient who died at the age of 79.
Significantly VOC (> 6 episodes/year) were seen in about 20% of cases, reflecting that recurrent VOC remains a marker of severe disease, and these patients accumulate disease-related end-organ damage that leads to their mortality [28]. However, about two-thirds of patients in this cohort (63.4%) had relatively mild disease, indicating the need for intense vigilance in picking up alarm signs for early mortality in some of these relatively mild disease cases.
Many of our patients had a history of ACS, stroke, hepatic & splenic sequestration as well as dactylitis, which were shown to be independent predictors of early childhood mortality in SCD patients [7–9,19]. Sixty-five (52.8%) patients had a prior history of ACS, one of the major complications of SCD, and contributed to the terminal event in 25(20.3%) of this cohort. Furthermore, 84 (68.3%) patients presented with fever, cough (24.4%), and 61(49.6%) had reduced O2 saturation, while 67 (54.5%) had abnormal findings in the chest examination in the terminal episode, with 82 (66.6%) needing ventilatory support. Similarly, 71 (57.7%) patients in this cohort had an abnormal chest X-ray, which was one of the defining features of ACS, and 77.8% had an abnormal CT scan of the chest. Importantly, Knight-Madden JM et al [29], and Tawfiq QA [16], et al, have shown that mechanical ventilation was found to predict mortality and increased utilization of hospital resources. In this cohort 66.6% needed ventilator assistance, the majority of these patients received simple blood transfusions [118 (95.9%)] and many needed exchange transfusions [115 (93.5%)] as well.
Sepsis seems to be a major contributor to the terminal event as seen in 49 (40%) of cases in this cohort. These patients were characterized by a significantly high CRP, along with a significant rise in the WBC counts from the baseline, and a significant drop in platelets. Among patients with sepsis, almost 80% of these patients had positive microbial cultures confirming sepsis. These patients received multiple broad-spectrum antibiotics during their terminal event. In the cooperative study of SCD from the USA, only a few bacteria were detected in lung samples at bronchoscopy, possibly since bronchoscopy was performed after empiric antibiotic administration, which is a universal practice [30]. Furthermore, even among SCD patients who had an autopsy done on them, sepsis was identified as the leading cause of death in a postmortem study [31]. More recent paper by Ballas SK indicated that sepsis contributed to death in a significant number of patients with SCD [32].
Among the miscellaneous causes, multiorgan failure with sudden death including cardiac events with dilated cardiomayopathy seem to be a leading cause, although pulmonary embolism, stroke and RTA were also other significant causes of mortality in this cohort. Sudden death with bone marrow emboli seems to be associated with HbSC and HbSβ+ Thalassaemia or the milder SCD phenotypes [33]. However, in this SCD patient cohort HbSβ+ Thalassaemia was seen in 22.5% with no HbSC patient and 75.5% patients having HbSS genotype. This is similar to findings from other studies, [15,18] Similarly pulmonary embolism (PE), stroke, and RTA were also significant causes of mortality in this cohort. PE is the leading presentation of thromboembolic complications in SCD, [34–35] and stroke prevalence was relatively low in our population (probably due to the high prevalence of alpha thalassemia, Alkindi et al [5]); however it remains as an important cause of morality. RTA contributed to about 5% of overall total deaths, although no data is available if it was linked to opioid use while driving, on not [32].
Laboratory data showed that there was a statistically significant drop in levels of hemoglobin and platelets count at the terminal event in this study cohort (Table 1). A low hemoglobin level has been previously shown to correlate with an increased risk of death with stroke [18]. Although the majority of patients had a baseline Hb > 7 g/dl, the median Hb fell from 9.5 to 6.8, during the terminal event, in all causes of death (Table 2). On the contrary, literature reports show that a higher hemoglobin (Hb) level correlated with an increased risk of acute chest syndrome and painful crisis [18]. This was also seen in this study, using the cutoff basal Hb > 7g/dl and basal WBC > 15 X109/L (Table 2). Consistent with this, many patients in this cohort needed either simple transfusion (95.9%) or exchange transfusions (93.5%).
In a previous prospective study by Plat. et al [7], the most straightforward laboratory risk factor related to the cause of death was fetal hemoglobin level. Patients with high levels had an improved life expectancy, while adults who had low levels of fetal hemoglobin as children were likely to die earlier than those who had high levels [7]. Thus, when using a cutoff of HbF of 8.5, there was a clear protective effect of high HbF against all types of death, with a significant proportion of these deaths occurring in the low HbF group, (p < 0.05, Wilcoxon signed ranks test, Table 1). However, there is heterogeneity of HbS-associated haplotypes in Oman. Although the Arab-Indian haplotype is common in the Eastern provinces in Saudi Arabia, [21] in Oman, it only constitutes 25% in the SCD patient population, while Benin and Bantu haplotypes form 50% and 25% respectively [36]. Another, confounding factor related to HbF is hydroxyurea use, which is seen in 36% of this current SCD cohort.
An elevated WBC count was an independent predictor of disease severity as seen in our study. Quinn CT & Miller ST [37] in a prospective study reported that leukocytes are known to be involved in the process of vaso-occlusion, and leukocytosis in adults is associated with an increased frequency of ACS and death [38]. Our patients also showed a significant rise in the WBC counts in the terminal event and using a cutoff of 15X109/L, it was shown that there was a statistically significant rise in the WBC counts with a median WBC count of 24 and IQR between 20 to 33 (Table 1) and this correlated significantly with sepsis during the terminal event.
High serum ferritin was thought to be one of the associated features of mortality among patients with SCD, reflecting both iron overload, as well as inflammatory status. Increased gastrointestinal absorption of iron has been reported in sickle cell disease, because of the associated chronic hemolysis, and anemia. In our study cohort, the median serum ferritin (ng/L) was 1064 with IQR between 323 to 2987. Interestingly, Akinbami et al [39], recently reported that 90% of subjects with sickle cell disease had normal iron stores.
Elevated CRP can occur in SCD, during steady state, and in crisis, as reported by us and Okocha et al [40–41]. It represents an underlying inflammatory/ infective, process or tissue necrosis and showed a statistically significant rise, in our cohort, especially among patients who died of sepsis/ multi-organ failure (Table 1). In our cohort, the basal median CRP was high (43 mg/L) and showed a statistically significant rise in the terminal episode with a median of 171 (p < 0.05). Among 67 cases with sepsis, 54 (79.1%) were culture-proven, with 11 (20.7%) patients, showing multiple microorganisms, while bacterial, fungal, and viral microorganisms were isolated in 32 (60.4%), 6 (11.3%), and in 4 (7.5%) respectively. The most frequent organisms were gram-negative organisms, similar to the recent report of the rising gram-negative organisms, reflecting an increasing use of external lines and acquisition of hospital acquired infections [42].
Our study demonstrated a high serum LDH, high serum bilirubin, and high CRP that further increased significantly toward the terminal event. Several studies have shown these observations and they indicate the degree of inflammation, tissue necrosis, infarction, and haemolysis that is accentuated during the terminal phase [29,43–44]. Our study demonstrated a statistically significant rise in these parameters in the terminal episode substantiating the above (Table 1). These factors with the associated leukocytosis, thrombocytopenia, and drop in hemoglobin heralded the onset of multiorgan failure in this cohort.
The cause of death was established by the course of illness during hospitalization, supportive laboratory evidence and telephonic enquiries with relatives in case of out of hospital deaths of these patients. Although ACS has remained a leading cause of death, there are no established risk factors for death in patients with SCD in Oman. In a study from Oman by Tawfic et al, [16] ACS was the main cause of ICU admission in patients with SCD. Further, the use of inotropic support and/or mechanical ventilation was an indicator of high mortality rate among these patients. Jaiyesimi O. & Kasem M,[17] showed that ACS was common irrespective of SCD severity in children from Oman, and all patients appeared to be at risk, but it was increased in patients with vaso-occlusive crisis.
When we looked at the time of death, we observed increased mortality due to ACS and sepsis during the colder months of the year (September to March). This is not unusual as viral infections such as respiratory syncytial virus and seasonal influenza predominate during that period and it is known to be one of the leading viral causes of ACS. Our own data and others showed similar patterns [45–46]. We also observed that over 30% of deaths occurred within 24h of hospital admission. This data is also similar to the previous studies and the highest numbers were reported in the ACS and sepsis groups [47]. In the absence of post-mortem studies, (which are rarely accepted by the patient’s relatives), the exact cause will remain elusive or putative, although the recent literature is suggestive of increased cardiovascular events in this situation.
SCD is a complex condition, with life-threatening complications that interfere with the patient’s normal life owing to repeated painful episodes, and the associated multi-system involvement. Although this is a retrospective study, however, it identified that ACS and sepsis were the most important preventable risk factor associated with mortality in this SCD patient cohort. A high index of suspicion and vigilance is required for patients presenting with fever, tachycardia, and signs of ACS, associated with a dropin Hb, with raised WBC, CRP, LDH, and ferritin, to identify these patients early and initiate appropriate rapid management protocols.
In summary, the ability to identify the risk factors that are associated with increased mortality among SCD patients permits accurate prognostication and provides effective prophylactic management strategies. This study shows that male gender, low HbF, substantial drop in hemoglobin and platelet, as well as increased in WBC counts, serum LDH, ferritin and CRP, correlated significantly with mortality risk during the terminal event in patients with SCD.