記住我
Abstract
Much has now been learned about the coronavirus disease 2019 (COVID-19) in the general population, but data for hemodialysis (HD) patients are limited. This is the first study of COVID-19 disease in patients undergoing maintenance HD in Pakistan. We studied the epidemiological, clinical, laboratory, radiological characteristics, and outcomes of a cohort of HD patients that contracted COVID-19 in our HD center from the first confirmed case on May 12, 2020, until September 9, 2020. Out of the total 423 patients being dialyzed in our center, 87 were suspected and 50 (11.8%) were confirmed for COVID-19. Male:Female ratio was nearly equal. The median age was 59.5 ± 9.99 years. Most patients developed mild disease. The most common symptoms were fever (82%). Ten (20%) had patchy bilateral opacity (ground-glass opacity) on the chest radiograph. Major complications were lymphocytopenia (36%), thrombocytopenia (30%), pneumonia (28%), and septic shock (6%). Eleven (22%) patients were hospitalized. Five required mechanical ventilation. Ten (20%) patients died. The relative risk of death with COVID-19 in HD patient was 1.46 with 95% confidence interval 1.15–1.84, (P = 0.003). The patients aged ≥60 years had 4.3 times more severe disease (P = 0.044) and died 3.3 times (P = 0.164) more than patients aged <60 years. HD patients have a high susceptibility to COVID-19 compared to the general population with an increased mortality rate and prolonged recovery time. Patients with age >60 years, female gender, diabetics, and those presented with more severe symptoms and laboratory parameters, had a higher fatal outcome.
Introduction 
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) a newly discovered virus causing what is now called coronavirus disease 2019 (COVID-19). It is a contagious disease which first emerged in Wuhan, Hubei Province, China, in December 2019,[1] and was declared a global pandemic by the World Health Organization (WHO) on March 11, 2020.[2] The disease has resulted in a large number of hospitalizations and intensive care unit (ICU) admissions, with now well-described pulmonary, cardiac, vascular, and renal complications.[3],[4],[5]
In Pakistan, the first case of COVID-19 was reported on February 26, 2020, from Karachi with an estimated populace of Pakistan as 204.65 million.[6],[7] According to Pakistan’s last update[7] on December 31, 2020, a total of 6,696,068 tests were done in Pakistan, 479,715 of which tested positive for COVID-19 (7.16%). Of the 479,715 cases, 435,073 patients recovered (90.7%), 2,206 remained critical (0.46%), and 10,105 died (case fatality rate: 2.1%). The coronavirus attack rate is estimated to be 234 per 100,000 Pakistani population.
Most studies published on COVID-19 mainly described the general epidemiological findings, clinical presentation, and clinical outcomes of patients in the general population.[8],[9],[10] According to the published literature, COVID-19 patients with underlying conditions such as diabetes, hypertension (HTN), cardiovascular disease (CVD), obesity, or the elderly are highly susceptible and often have a more serious disease.[5] However, there are only very minimal data available on the impact of COVID-19 epidemic on clinical characteristics or outcomes of patients on maintenance hemodialysis (HD), and as far as we know, none from Pakistan.
Pakistan is a developing country with limited financial resources as compared to developed countries to combat with COVID-19 outbreak. Even with fewer resources, Pakistan has taken rigorous measures such as opening many field and specialist hospitals, laboratories for testing, quarantine facilities, awareness campaigns, and lockdown to control the spread of virus to prevent and combat this deadly disease.[11] However, considering the large population size of HD patients,[12],[13] the compromised immune function of uremic patient[14] along with the increased frequency of comorbidities such as diabetes mellitus (DM), HTN, and CVD among maintenance HD patients, which tend to make them more susceptible not only to COVID-19 infection but also to other severe illness, HD patients need special consideration.
Maintenance HD patients at HD center, additionally come in close contact with health-care workers and other patients in a relatively confined environment for a considerably long time making it more challenging for health-care authorities to prevent spread, and manage infectious diseases in these patients than in general population,[15],[16] especially in resource-limited countries like Pakistan. In addition, there are issues related to difficulty in early diagnosis, preventing cross-infection, disposal of spent dialysate and adjustment of medications.
Therefore, it would be worth knowing whether the clinical course of patients undergoing maintenance HD having COVID-19 is different from that of COVID-19 patients in general.
In this study, we aimed to review epidemiological, clinical, laboratory, and radiological characteristics and outcomes of a cluster of HD patients who contracted COVID-19 in our HD center from the first confirmed case COVID-19 on May 12, 2020, until September 9, 2020. We expect our findings to shed light on the appropriate management of the HD center and HD patients in the face of COVID-19 or when other similar epidemics emerge.
Methods This observational cohort study was conducted at the Department of Nephrology at The Kidney Centre Postgraduate Training Institute (TKC-PGTI) after approval by the institutional ethical review committee. TKC-PGTI is a tertiary-level renal care facility with dialysis a unit that accommodates 66 HD patients at a time and performs 82,500 dialysis sessions annually, following standards defined by the European Best Practice Guidelines. Nearly all long-term maintenance HD patients are dialyzed for 4 h three times a week.
We reviewed the epidemiological, clinical, laboratory, radiological characteristics, and outcomes of a cohort of HD patients that contracted COVID-19 in our HD center from the first confirmed case on May 12, 2020, until September 9, 2020. A total of 50 confirmed cases of COVID-19 were studied.
Data collection
Information collected included demographics data, exposure history, medical history, dialysis vintage, comorbidities, symptoms, signs, radiologic, laboratory tests, complications, and treatment received during COVID-19 infection and outcome. Informed consent was taken from the patient and for those that had expired; their next of kin were approached through telephone after explaining the purpose of the study. Data were also collected retrospectively from the medical records; however, if data were missing from the records or clarification was needed, data were obtained by direct communication with patients or their families, attending doctors, and other health care providers.
Laboratory assessments consisted of a complete blood count, C-reactive protein, ferritin, alanine aminotransferase, albumin, and intact parathyroid hormone. Radiologic abnormality was determined on the basis of the description on medical charts or in the documentation. Treatment that these patients received from their primary healthcare provider included oxygen, antibiotic therapy, hydroxychloroquine, remdisivir, corticosteroid therapy, and respiratory support. Complications studied were lymphocytopenia, thrombocytopenia, pneumonia, acute hepatic injury, septic shock, disseminated intravascular coagulation (DIC), and acute respiratory distress syndrome (ARDS). The outcomes were hospitalization, the use of mechanical ventilation, recovery, or death.
Study definitions
A confirmed case of COVID-19 was defined as a positive result on at least one of two tests done 24-h apart on real-time reverse transcription-polymerase chain reaction (RT-PCR) assay of nasal swab specimens.[17] A systematic review of the accuracy of COVID-19 tests reported false-negative rates of between 2% and 29% (equating to sensitivity of 71%–98%), based on negative RT-PCR tests which were positive on repeat testing.[18] As per institutional policy, we needed at least two negative tests to declare COVID-19-infected patients unable to transmit virus to others.
The date of disease onset was defined as the date when the symptoms were first noticed. The incubation period was defined as the interval between the earliest date of likely contact with the transmission source (person with suspected or confirmed case) and the earliest date of symptom onset (i.e., cough, fever, fatigue, or myalgia). Laboratory tests were defined as the first test results available after confirmation of COVID-19 diagnosis. Lymphocytopenia was defined as a lymphocyte count of less than 1500 cells/mm3. Thrombocytopenia was defined as a platelet count of less than 150,000/mm3. Pneumonia was diagnosed using the American Thoracic Society guidelines for community-acquired pneumonia,[19] septic shock, using the third international consensus definition,[20] DIC according to the International Society on Thrombosis and Hemostasis criteria,[21] and ARDS was defined according to the Berlin definition.[22] Acute hepatic injury was defined as an elevation in alanine aminotransferase of more than 10 times the upper limit of normal.[23]
We defined the degree of severity of COVID-19 as mild, moderate–severe, and critical. Mild refers to patients who had mild symptoms without manifestation of viral pneumonia on chest X-ray. Moderate–severe refers to patients who had symptoms such as fever and respiratory tract symptoms with features of viral pneumonia on chest X-ray with or without respiratory rate ≥30 breaths/min; oxygen saturation ≤93% at rest state, and pulmonary lesion progression of ≥50% within 24–48 h on radiologic imaging. Critical refers to patients with respiratory failure requiring mechanical ventilation, and/or with the presence of shock, and/or another organ failure that requires monitoring and treatment in the high-dependency unit (HDU) or ICU.
Statistical Analysis Data were entered and analyzed by IBM SPSS Statistics software version 21.0 (IBM CORP., Armonk, NY, USA). Mean with standard deviation and median with interquartile ratio were calculated for continuous variables, while for categorical variables, frequencies with percentage were obtained. To observe the effect of different variables on the severity of disease and the outcome of COVID-19, logistic regression was run and the odds ratio with 95% confidence interval (CI) was obtained. For binary logistic regression, we merged the three categories of disease severity into two categories as severe and nonsevere. The relative risk of death was compared with deaths during the same period in 2019. P ≤0.05 was considered statistically significant.
Results A total of 423 maintenance HD patients were registered at the TKC-PGTI, during the study. Out of these, 87 were suspected and 50 patients (11.8%) were confirmed as COVID-19 on RT-PCR assay of nasal swab specimens. We studied the clinical characteristics and outcomes of these 50 patients, who contracted COVID-19 in our HD center from the first confirmed case on May 12, 2020, until September 9, 2020.
There were 26 (52%) males and 24 (48%) females. The median age of the patients was 59.5 ± 9.99 years and 60% of the patients were above 60 years of age. The major causes of CKD were DM in 22 (44%) followed by HTN in 11 (22%). The mean HD vintage was 4.59 ± 4.7 years. All patients had one or more coexisting morbidities. The most common comorbidity was HTN (94%) followed by diabetes (44%) and coronary artery disease (CAD) (36%). Fifteen patients (30%) had a history of former or current smoking. Exposure history was positive in eight (16%) of patients, two (4%) patients had a travel history to high prevalent COVID-19-infected areas within the country.
Three patients after their initial symptoms remained asymptomatic throughout the course. The median incubation period was five days. Thirty-four (68%) patients developed mild disease, six (12%) moderate–severe, whereas 10 (20%) were critical. The most common symptoms were fever in 41 (82%) followed by fatigue in 32 (64%), cough in 19 (38%), and myalgia/arthralgia in 15 (30%).
Out of 50 patients, 12 (24%) revealed abnormal findings on chest X-ray. The most common pattern was patchy bilateral opacity (ground-glass opacity), found in 10 (20%) patients. Mean hemoglobin level was 10.2 ± 1.4 g/dL, leukocytes 7.8 ± 4.1/mm3, and platelets 217 ± 106/mm3. Patients with severe disease had worse laboratory abnormalities than nonsevere. Detailed demographic, clinical, laboratory, and radiological findings of study participants are given in [Table 1].
Table 1: Demographic, clinical, laboratorial, and radiological findings of hemodialysis patients with coronavirus disease 2019 according to disease severity.The major complication was pneumonia in 14 (28%) patients, followed by septic shock in three (6%) patients. Lymphocytopenia and thrombocytopenia were present in 18 (36%) and 15 (30%) of patients, respectively. Oxygen and intravenous antibiotics were given in 10 (20%) patients, four (8%) of patients received systemic steroids. Eleven (22%) patients needed hospitalization with three (6%) in HDU and six (12%) in ICU. Mechanical ventilation was needed in five critical patients. Ten (20%) patients died [Table 2]. Survival at week 1 from the date of a positive SARS-CoV-2 test was 98% and survival at week 2 was 86%. The mean time of death from the date of positive SARS-CoV-2 PCR test was 11.9 ± 5.8 days and from the date of onset of symptoms was 15.2 ± 4.96 days. The mean time of recovery was 30.6 + 14.0 days from the date of positive test, whereas 34.4 + 14.6 days from the date of onset of symptoms.
Table 2: Complications, treatment used, and outcome of dialysis patients with coronavirus diseases 2019 (n=50).Patients aged 60 years or more had 4.3 times more severe disease than patients with age <60 years (P = 0.044). Males had 2.25 times more severe disease than females, although it was not statistically significant (P = 0.169). Comorbidities such as DM, HTN, CAD, and obesity did not significantly relate to the severity of the disease. The patients who were vaccinated with flu vaccine suffered 3.6 times more with severe disease as compared to patients without vaccination (P = 0.049), while BCG vaccination did not effect on disease severity in our study patients [Table 3].
Table 3: Association of study variables with disease severity and outcome.The patient 60 years or more had 3.3 times more deaths as compared to patients less than 60 years of age, although it was not statistically significant (P = 0.164). Among all comorbidities, patients with DM had 3.1 times more deaths than nondiabetic, but this was again not statistically significant (P = 0.13). Patients vaccinated with the flu vaccine had 3.2 times more deaths than non-vaccinated (P=0.131). Females had 2.3 times more deaths than males (P = 0.131) [Table 3].
The relative risk of death associated with COVID-19 in HD patients was found to be 1.46 with 95% CI 1.15–1.84, (P = 0.003) [Table 4].
Table 4: Relative risk of death associated with COVID-19 in hemodialysis patient. Discussion The sudden arrival of the COVID-19 pandemic raised an unforeseen challenge for the health community. The disease has resulted in a large number of hospitalizations and ICU admissions, with now well-described pulmonary, cardiac, vascular, and renal complications.[3],[4],[5] Much is now known about the disease in the general population, but there are still limited data available for HD patients, especially in the developing countries like Pakistan. This is the first study on patients undergoing maintenance HD with COVID-19 disease from Pakistan.
Out of our 423 registered HD patients, 87 patients were suspected and 50 patients (11.8%), were confirmed to have COVID-19 infection which suggests high susceptibility of HD patients to COVID-19 as compared to the general population in Pakistan where, out of the tested population, only 7.16% were positive. The coronavirus attack rate is estimated to be 234 per 100,000 Pakistani population as of December 31, 2020.[7] Similar observations were made in studies from China[24],[25] and Italy.[26] Various likely explanations can be the compromised immune function of uremic patients,[14] and increased frequency of comorbidities in HD patients. Kwan et al[27] reported that dialysis patients have a higher rate of contracting SARS (1.72%) than the general population (0.003%) in Hong Kong. Given that SARS-CoV-2 has up to 85% sequence similarity with SARS,[1] there is the possibility that the epidemiologic and clinical characteristics of COVID-19 could resemble that of SARS in patients undergoing maintenance HD.
The physical distancing of these patients during thrice-weekly HD sessions, in the waiting room, and during transport to HD facilities is extremely challenging. In addition, these patients come in contact with healthcare providers’ multiple times a week. Even though in our dialysis center, all patients are checked for temperature twice, and before starting their session’s patients were asked about symptoms related to COVID-19. Those who were suspected were then dialyzed in a separate area at different times from the usual shift to minimize interaction with other patients. These suspected and subsequently confirmed patients had to have two negative PCR negative tests before they could go back to their original days and times. All COVID-19-confirmed patients were dialyzed in separate area and on different days. Even after all these measures, preventing cross-contamination remained highly challenging. In our study of patients, the history of exposure was positive in only eight (16%) cases, only two (4%) patients had a travel history to the high prevalence area within the country. Three (6%) patients after their initial symptoms remained asymptomatic. The mean incubation period was five days. This may suggest that the primary source of COVID-19 spread maybe asymptomatic patients or patients in incubation period,[28],[29] or close contact with many individuals in the HD center.
Approximately the same number of male (26) and female (24) HD patients had COVID-19 in our study population. Male patients acquired severe disease 2.25 times more than females (P = 0.169), but females had 2.3 times more deaths than males secondary to COVID-19. Our finding contradicts the popular belief for sexual dimorphism in innate immune responses to infectious organisms,[30] which suggests that in general, females generate more robust and potentially protective humoral and cell-mediated immune responses following antigenic challenge than their male counterparts. A plausible explanation for this observation could be that in Pakistan, in general, gender inequality leads to a vulnerable situation for women with unequal power relationships in a family and the social and economic status of women within the community. This gender inequality leads to an increase frequency of undernutrition and anemia in Pakistani women.[31] HD itself is a catabolic state, and several studies have shown the role of gender in predicting malnutrition, with an increased rate of malnutrition seen in females.[32],[33] In our study, there was a difference in mean hemoglobin (10.5 ± 1.43 vs. 9.91 ± 1.33, P = 0.124) and albumin (3.48 ± 0.46 vs. 3.36 ± 0.53, P = 0.382) level, between male and female participants. Although not statistically significant, this general malnutrition and frail health status of females may predispose them to fatal outcome in the face of COVID-19 infection.
Our data also showed that the majority of HD patients who contracted COVID-19 were older with a mean age of 59.5 ± 10 (median 60.5) with 60% above 60 years of age, which correlates with the findings of studies from China[24] and Italy[26] in HD patients but significantly higher than general population (median 47 years) data by Guan et al[9] from China. However, in data by Guan et al, patients with primary composite endpoint (admission in ICU, underwent mechanical ventilation, or died) were of older age (median 63 years). Our study also showed that the patients aged >60 years have 4.3 times more incidence of severe disease (P = 0.044), which was statistically significant, and died 3.3 times more as compared to patients less than 60 years of age, although it is not statistically significant (P = 0.0164). This might be because of small sample size. This correlates with the WHO report[34] and a study from China[8] that found that adults over 65 years of age represent 80% of hospitalizations and have a 23-fold greater risk of death than those under 65. Comorbidities such as CVD, diabetes, and obesity associated with older age, may increase the chances of fatal disease, but these alone do not explain why age is an independent risk factor. Mueller et al[35] tried to explain molecular differences between young, middle-aged, and older individuals that may explain why COVID-19 is a mild illness in some but life-threatening in others. However, why the disease is perilous in older people is not yet known.
Along with age, diabetes was suggested by some studies[36],[37] as a possible predisposing factor. The result of our study shows that diabetes (44%) was the second most common comorbidity after HTN (94%), however, univariate analysis did not find a statistically significant relationship between diabetes and disease severity (P = 0.425). Patients with diabetes died 3.1 times more than non-diabetes, however, this did not reach statistical significance (P = 0.13) possibly because of small sample size. Similarly, other comorbidities such as HTN, CAD, and obesity did not significantly relate to disease severity or death in our study. Smoking, HD vintage, and type of dialyzer use showed no significant relation with disease severity or outcome.
Our study found that the patients who were vaccinated with the flu vaccine suffered 3.6 times more with severe disease as compared to patients without vaccination (P = 0.049), and died 3.2 times more than nonvaccinated (P = 0.131). This result again contradicts the result of studies from Italy,[38] and earlier scientific research belief[39] that higher influenza vaccination rates were associated with fewer deaths from COVID-19 in elderly or hospital workers who got vaccinated were significantly less likely to develop COVID than those who did not. Later, the National Kidney Foundation did explain that getting the flu vaccine will not protect against COVID-19.[40] This negative relation between the flu vaccine and frequency of disease severity and death in our study participants is still inexplicable.
In contrast to initial numerous epidemiological analyses that showed a correlation between the incidence of COVID-19 and BCG vaccination policies,[41],[42] our study supports the observation by Hensel et al[43] that after correction for confounding variables, there was no association between BCG vaccination policy and COVD-19 spread rate or percentage mortality. Similarly, our study found no relation between BCG vaccination and disease severity or death in our study participants.
The most common symptoms were fever in 41 (82%) followed by fatigue in 32 (64%), cough in 19 (38%), myalgia/arthralgia in 15 (30%). The disease symptoms of our patients were similar to those of the general population[9] and study by Wang et al.[8] Some studies in HD patients[24],[26],[37] had less frequency of fever than the general population; however, the frequency of fever in their study population also increased with disease severity or in hospitalized patients.
The most common radiological finding on chest X-ray in our patient was bilateral patchy bilateral opacity (ground-glass opacity) which correlates with the finding in other studies.[8],[24],[26] Financial constraints in resource-limited countries like Pakistan limited the accessibility of CT chest in these patients, however correct diagnosis of COVID-19 with chest X-rays ranged from 57% to 89% which is compatible with chest CT that correctly diagnosed COVID-19 in 89.9%.[44]
Lymphocytopenia and thrombocytopenia were present in 18 (36%) and 15 (30%) patients, respectively, however, contrary to findings in other studies,[10],[24] we did not find a trend of decline in severe or critical patients. Other complications were pneumonia in 14 (28%) patients followed by septic shock and acute hepatic injury, each in three (6%) of patients. The majority of complications occurred in patients with severe to critical disease.
Our institute did not treat these COVID-19 patients, therefore some of the data regarding the treatment of severe to critical patients who were referred to COVID-19 specialized centers may be missing. Although we did obtain their data by direct communication with patients or their families and their attending doctors, some of the information may not be properly known or remembered by their relatives or attending doctors. Based on available data, apart from oxygen therapy and intravenous antibiotics that were received by 10 (20%) patients, only four (8%) patients received systemic glucocorticoids, hydroxychloroquine, and remdisivir. The majority of patients did not receive treatment that claimed to be beneficial for COVID-19 by some studies.[45],[46] This may be due to resource limitations along with the absence of randomized controlled trails to support different therapy for COVID-19 in patients on MHD and low certainty of the efficacy of these treatments.
Out of 50 patients in our study, 34 (68%) had mild disease, and only one required hospitalization. Out of 16 patients, who developed moderate-severe to critical disease, 10 patients were hospitalized. Six patients were treated in ICU, of which five required mechanical ventilation. Patients with age >60 years, female gender, diabetes, those presented with more severe symptoms and worse laboratory parameters, had a more fatal outcome.
A high rate of death (20%) was found in HD patients with COVID-19 in our study, this is similar to the study finding from Spain[47] and the US.[36] This is much higher than that observed in the general population in Pakistan (2.1% as that of December 31, 2020).[7] One patient died at the end of 1st week from the date of a positive SARS-CoV-2 PCR test, while by the end of 2nd week, seven patients died, that gives survival at week 1 from the date of a positive SARS-CoV-2 test of 98% and survival at week 2 of 86%. The finding of high mortality rates among critical patients were also similar to studies from Italy[48] and US.[36]
We compared the number of deaths in 2019 in our dialysis center, with the number of deaths during the same period in 2020, and studied the impact of COVID-19 on the mortality in HD patients. The relative risk of death associated with COVID-19 in HD patients was 1.46 with 95% CI 1.15–1.84, (P = 0.003). Recovery time in our study patients was longer i.e. 30.6 ± 14.0 days from the date of positive test whereas 34.4 ± 14.6 days from the date of onset of symptoms which is even more prolonged in patients with moderate–severe disease (45.8 ± 17.49 and 52.5 ± 20.1 respectively). This finding correlates with the report from Hubei province, China.[49] A possible explanation given was that during the outbreak of SARS, it was observed that the duration of shedding virus through stool or breath was significantly longer in HD patients than in other patients.[27] Considering the high biological similarity between SARS-CoV-2 and SARS-CoV, HD patients infected with SARS-CoV-2 may also take longer to clear the virus, and these patients may need a longer quarantine period to prevent the spread of infection.
Our study has certain limitations, some of which have already been discussed above at relevant places that include a small sample size, nonavailability of CT chest, possible missing of some information, especially about the management of severe to critical patients. Some of the laboratory tests as LDH and D-dimer which are considered a marker for a prognosis for the COVID-19 patient in the general population were not checked in our study because of the unavailability of studies to support their role with certainty at the time when our study enrollment was started. Later, some studies claim that D-dimer levels are probably not good predictors of outcome in maintenance HD patients as its elevation has been described in maintenance HD patients in stable conditions too, and there is evidence of D-dimer clearance with dialysis therapy that can affect its concentration independent of disease state.[47]
Despite these limitations, this is the first a study from Pakistan or any developing country providing information about epidemiologic and clinical characteristics and outcome of patients undergoing maintenance HD with COVID-19. Some of the important findings reported in our study will help in understanding the clinical characteristics and outcome of COVID-19 in maintenance HD patient. They will be helpful to make policies regarding intervention measures that should be taken by maintaining HD centers effectively for early diagnosis, preventing cross-infection, upgrading the level of prevention for staff, strict monitoring of suspected patients, and timely referral to specialized COVID-19 care centers.
Conclusion Our study found that there is a high susceptibility of HD patients to COVID-19 as compared to the general population with asymptomatic patients, and patients in the incubation period may be the primary source of COVID-19 spread within HD center. Disease symptoms, radiological findings, and laboratory tests were found similar to that of the general population. Patients with age >60 years, female gender, diabetes, those presented with more severe symptoms and laboratory parameters, had more fatal outcome. Patient vaccinated with the flu vaccine had more severe disease and died more than nonvaccinated, while no such association was found with BCG vaccination. Mortality in HD patients was higher than general population especially with critical disease with prolonged recovery time among survivals. The relative risk of death associated with COVID-19 in HD patients was found to be 1.46 times.
Acknowledgment
We would also like to extend our gratitude to Dr. Bina Salman for her help with the statistics portion of this paper.Conflict of interest: None declared.
References 
留言 (0)