Circumstances of Initiation of Hemodialysis: Importance of Nephrological Follow-Up ()
1. Introduction
End-stage renal disease (ESRD) is a progressive and irreversible loss of renal function, resulting in a fall in glomerular filtration rate (GFR) to below 15 ml/min/1.73 m2. It is a major public health problem due to its severity and its medico-economic impact on the healthcare system. In most cases, chronic kidney disease (CKD) remains asymptomatic until it reaches an advanced stage.
The timing of the initiation of renal replacement therapy depends not only on the level of residual organ function (which is estimated by glomerular filtration rate), but also on the patient’s clinical condition and local medical practices.
Hemodialysis is most often used on an emergency basis, as the diagnosis of CKD is usually made late. In the end stage, this pathology requires long-term renal replacement therapy, either through extra-renal purification (peritoneal dialysis or hemodialysis) or through renal transplantation or grafting. Hemodialysis is the most widespread and accessible method of renal replacement. It is a detoxification method based on blood filtration through an artificial membrane [1] [2]. Starting dialysis on a planned basis remains a challenge for many patients. Previous studies have shown that late patient referral (LR) to a nephrologist is associated with emergency dialysis, and that patients starting emergency dialysis are more prone to severe comorbidities [3]. According to many authors, this LR to the nephrologist is responsible for the increased mortality of dialysis patients, in association, of course, with other known causes: advanced age, the severity of comorbidities (notably cardiovascular), uremia-induced impairment of immune defenses and nutritional status... [4] [5]. Early referral (ER) with early management (by the nephrologist) of kidney disease and the factors associated with its progression can delay initiation of replacement therapy, reduce the frequency of emergency dialysis, improve survival after dialysis and better prepare for the possibility of renal transplantation [6].
In North America, in a multicenter study published in 2014, D. C. Crews et al. [7] noted that 35.4% of patients had benefited from late initiation of dialysis.
In France, P. Tuppin et al. [6] in 2017 reported that 33% of their patients started hemodialysis as emergencies.
In Senegal, in a single-center study of emergency hemodialysis carried out in 2020, Diawara et al. [8] reported a prevalence of 40.68% in the nephrology-hemodialysis department of Centre Hospitalier de Thiès.
In Guinea, in a single-center study of emergency hemodialysis published in 2023, F. Diakité et al. reported a prevalence of 59.16% at “Donka National Hemodialysis Center”, with a mortality rate of 16.9% [9].
Given the difficulties associated with emergency dialysis, as well as the high morbidity and mortality rates among patients undergoing this type of dialysis in our country, we undertook this study with the aim of elucidating the conditions under which dialysis is initiated at Donka National Hemodialysis Center, and then to compare the advantages of early versus late nephrological follow-up for the evolution of hemodialysis patients.
It should be noted that the value of this study lies in the fact that it would raise awareness of the need for early specialized follow-up of CKD patients in our sub-Saharan African regions, where the frequency and quality of dialysis are already poor. As we have shown above with data from the literature, the resulting corrective measures would probably reduce morbidity and mortality in hemodialysis.
2. Patients and Methods
This was a prospective descriptive and analytical study lasting 3 months, from September 01 to November 30 2022, carried out in the Hemodialysis Center at Donka National Hospital.
The target population consisted of all patients with renal failure, regardless of type or stage. These patients were either already on hemodialysis, hospitalized or being followed as outpatients.
All patients with CKD who had been followed in the department for at least six months prior to planned hemodialysis were included. Also included were patients with confirmed evoluted CKD who had been followed for less than six months, or who were not known to the department at all, and who were taken for emergency hemodialysis in the presence of one or more indications.
Not included were all inpatients or outpatients with a diagnosis other than renal failure, those diagnosed with acute renal failure, and CKD cases not yet at the hemodialysis stage during the study period.
The recruitment method was characterized by an exhaustive consideration of all patients with advanced CKD followed in the department, those on hemodialysis, all meeting our selection criteria explained above. A pre-established survey form was used to collect the necessary data from patients. Qualitative variables were expressed as proportions and quantitative variables as means ± standard deviation.
Quantitative variables included: age, biological parameters such as hemoglobin, leukocyte and platelet levels, creatinine levels, creatinine clearance, calcium levels, natremia and kalemia.
Qualitative variables included: gender, pathological history (Arterial hypertension, diabetes, phytotherapy, urinary tract infection), patient’s general condition (divided into unsatisfactory, impaired and coma), indications for dialysis (acute lung edema, hyperkalemia, uremic poisoning, metabolic acidosis, uremic pericarditis, coma) and finally initial nephropathy (Chronic glomerulonephritis, diabetic, vascular, tubulointerstitial and undetermined nephropathy).
The following definitions have been adopted for these expressions:
-Emergency hemodialysis: Or unscheduled hemodialysis, necessary when the patient’s vital prognosis is threatened (acute lung edema, threatening hyperkalemia, metabolic acidosis, etc.).
-The opposite of this expression is programmed hemodialysis.
-Early referral: Referral of patients with CKD to the Donka University Hospital Nephrology Department at least 6 months before the start of dialysis.
-Late referral: Late referral of patients with CKD to the Donka UHC Nephrology Department, less than 6 months before the start of dialysis; again emergency referral of a patient never known to the department, requiring immediate dialysis treatment.
Data entry and the design of tables and figures were carried out using Word and Excel software from Pack Office 2019. Results were analyzed using Epi-info software version 7.2.3.1. The Chi-square statistical test was used to compare proportions, and the p value was considered statistically significant when less than 0.05.
The information obtained was used for purely scientific purposes, and the confidentiality of each patient was respected.
3. Results
Of a selection of 78 cases of renal failure starting hemodialysis during the study period, 70 had CKD, i.e., 89.7%.
The sex ratio M/F was 0.7 in favour of women, who accounted for 58.6% of cases.
Among patients initiated on hemodialysis during our study period, the most common age group was 50 - 60 (27.1%), followed by 28 - 38 (24.3%). The mean age of the patients was 46.4 ± 14.7 years, with extremes of 17 and 75 years.
Hypertension was by far the most common medical history, at 84.3% of cases.
Main clinical signs on admission: anemia syndrome in 100% of cases, uremic syndrome (74.3%) and peripheral edema (20%).
About general condition: 47.1% had altered general condition, 41.4% had unsatisfactory general condition and 11.4% presented disorders of consciousness ranging from obnubilation to coma.
Mean kalaemia was 5.462 ± 1.512 mmol/l.
Mean creatinine clearance was 5.475 ± 3.368 ml/min/1.73m2.
In terms of hemoglobin level, 91.8% had a level < 10 g/dl, 12.5% had between 5 and 6.9 g/dl.
Initial kidney disease: chronic glomerulonephritis was the main cause of CKD (45.7%).
The main indications for hemodialysis: uremic intoxication (74.3%), hyperkalemia (31.4%).
Initial vascular approach: femoral catheterization (75.7%), simple jugular catheterization (15.7%), tunnelled jugular catheterization (8.6%).
Of 57 late referrals: temporary catheterization = dominant vascular access (54 cases), mainly femoral catheterization (84.2%) because patients came in with life-threatening complications. Here, only 3 cases of tunneling.
On 13 early reference cases (generally more stable patients): the difference between permanent and temporary vascular approaches could be demonstrated by the following percentages: 76.9% vs 23.1%.
Bivariate analysis:
Diabetes history (p value = 0.02): more in the late than early group 7 (58.3%) vs 5 (41.7%);
Uremic intoxication (p value = 0.02): more frequent in the late group 39 (75.0%);
Hypercreatininaemia was statistically significant (p = 0.02), much higher in the early reference group (198.270 ± 101.230 mg/L).
Temporary vascular approach: femoral catheterization (p value = 0.001) and simple jugular catheterization (p value = 0.013).
Permanent vascular approach: tunnel catheterization (p value = 0.03), with 3 cases of late referral and 3 cases of early referral.
The late reference recorded deaths (4 cases) during the first 3 months of dialysis, although there was no statistical link between the reference mode and mortality (p = 2).
4. Discussion
We conducted a prospective descriptive and analytical study lasting three (3) months, from September 01 to November 30, 2022. We collected 78 cases of renal failure initiated on hemodialysis during the study period. Among them, eight (8) were suffering from acute renal failure of various types, and therefore not to be included in the study; 70 were CKD patients, i.e., a frequency of 89.7%. These were the subjects of our study (Figure 1).
The M/F sex ratio of our sample was 0.7 in favor of females, who accounted for 41 subjects, or 58.6% of cases (Figure 2). This result differs from that reported by S. Cheikh et al. who, in their work published in 2021 on 371 patients admitted to nephrology for emergency hemodialysis at Ibn Rochd University Hospital in Casablanca, noted a M/F sex ratio of 1.16 [10]. We found no explanation for this difference in gender predominance.
The mean age of the patients was 46.4 ± 14.7 years, with extremes of 17 and 75 years (Table 1). Our result is lower than that of T. Mesbahi et al. [11] in Tunisia in 2020, who, in their study of 229 chronic renal failure patients initiated on hemodialysis, noted a mean age of 60.2 ± 15.3 years.
Figure 1. Flow chart of early and late follow-up in CKD patients initiated on hemodialysis.
Figure 2. Gender distribution of 70 CKD patients initiated on hemodialysis.
Table 1. Age distribution of 70 CKD patients initiated on hemodialysis.
Age Brackets (in years) |
Numbers (N = 70) |
Percentages (%) |
17 - 27 |
7 |
10 |
28 - 38 |
17 |
24.3 |
39 - 49 |
15 |
21.4 |
50 - 60 |
19 |
27.1 |
61 - 71 |
10 |
14.3 |
72 - 80 |
2 |
2.9 |
Total |
70 |
100 |
Mean age: 46.4 ± 14.7 years; Extremes: 17 years and 75 years.
Hypertension was the most common CKD comorbidity, accounting for 84.3% of cases (Table 2). Our results are similar to those of M. T. Yawovi et al. [12] in Togo in 2021, who found that hypertension and diabetes were present in 61.6% and 44% of their patients respectively. These results corroborate data from the literature, which reports that hypertension and diabetes are traditionally dominant risk factors for CKD [13].
Table 2. Breakdown of 70 CKD patients initiated on hemodialysis, by medical history.
History |
Numbers |
Percentages (%) |
High blood pressure |
59 |
84.3 |
Diabetes |
12 |
17.1 |
Phytotherapy/NSAIDs |
11 |
15.7 |
Recurrent urinary tract infection |
1 |
1.4 |
Clinical signs on admission were dominated by anemia syndrome in 100% of cases, followed by uremic syndrome (74.3%) and peripheral edema (20%). Consequently, 47.1% of our patients had an altered general condition, 41.4% had an unsatisfactory general condition and 11.4% presented disorders of consciousness ranging from obnubilation to coma (Table 3). Our results are not far from those of D. E. Sinomono et al. [14] in their study on “Initiation of dialysis in Brazzaville” published in 2018 who had found anemia in 91.4% of their patients, uremic syndrome in 72.7% and hydrosodium overload in 2.4%.
Table 3. Comparison of data between monitored and unmonitored CKD patients at the start of hemodialysis.
|
Initiation |
p value |
Late (N = 57) |
Early (N = 13) |
Gender |
-Male |
24 (82.8%) |
5 (17.2%) |
0.53 |
-Female |
33 (82.8%) |
8 (19.5%) |
|
-Average age |
45.7 ± 15.1 |
49.5 ± 13.1 |
0.32 |
Pathological history |
-Arterial Hypertension |
47 (79.7%) |
12 (20.3%) |
0.37 |
-Diabetes |
7 (58.3%) |
5 (41.7%) |
0.02 |
-Phytotherapy |
10 (90.9%) |
1 (9.1%) |
0.37 |
-Urinary tract infection |
1 (100%) |
0 (0%) |
0.81 |
General condition |
-Unsatisfactory general condition |
22 (75.9%) |
7 (24.1%) |
0.31 |
-Impaired general condition |
27 (81.8%) |
6 (18.6%) |
0.93 |
-Coma |
8 (100%) |
0 (0%) |
0.34 |
Hemodialysis indications |
-Acute lung edema |
5 (100%) |
0 (0%) |
0.26 |
-Hyperkalemia |
18 (81.8%) |
4 (18.2) |
0.95 |
-Uremic poisoning |
39 (75.0%) |
13 (25.0%) |
0.02 |
-Metabolic acidosis |
3 (75.0%) |
1 (25.0%) |
0.56 |
-Uremic pericarditis |
2 (100%) |
0 (0%) |
0.66 |
-Coma |
8 (100%) |
0 (0%) |
0.34 |
Biology (mean value) |
-Hemoglobin level (g/dl) |
8.8 ± 7.6 |
8.5 ± 1.6 |
0.756 |
-Creatininemia (mg/l) |
134.2 ± 62.6 |
198.3 ± 101.2 |
0.02 |
-Urea (mmol/l) |
139.4 ± 92.9 |
34.3 ± 23.0 |
0.22 |
-GFR (ml/min/1.73m3) |
5.9 ± 3.4 |
3.7 ± 3.0 |
0.095 |
-Calcium (mg/dl) |
8.1 ± 1.5 |
8.3 ± 1.1 |
0.747 |
-Sodium (mEq/l) |
132.7 ± 10.2 |
133.1 ± 4.1 |
0.92 |
-Potassium (mEq/l) |
5.4 ± 1.5 |
5.9 ± 1.7 |
0.38 |
-White blood cells (103/ml) |
8980.5 ± 7319.3 |
7343.3 ± 4380.5 |
0.71 |
-Platelets (103/ml) |
270522.0 ± 210638.3 |
272000 |
0.99 |
Initial nephropathy |
-Diabetic nephropathy |
5 (62.5%) |
3 (37.5%) |
0.14 |
-Vascular nephropathy |
19 (86.4%) |
3 (13.6%) |
0.47 |
-Chronic glomerular nephropathy |
29 (80.6%) |
7 (19.4%) |
0.84 |
-Tubulointerstitial nephropathy |
6 (85.7%) |
1 (14.3%) |
0.75 |
-Undetermined nephropathy |
0 (0%) |
1 (100%) |
0.18 |
Vascular approach |
-Femoral catheter |
48 (90.6%) |
5 (9.4%) |
0.001 |
-Jugular catheter |
6 (54.5%) |
5 (45.5%) |
0.013 |
-Tunneled catheter |
3 (50%) |
3 (50%) |
0.03 |
All this points to the late admission of patients with complications. In many African countries, the diagnosis of CKD is often made at the terminal stage, with access to dialysis being difficult due to its high cost and the limited presence of dialysis centers. This is also in line with literature data explaining complications in all organs and organ systems as renal function deterioration progresses [14] [15].
Mean kalemia was 5.462 ± 1.512 mmol/l (Table 3). In the work of S. Cheikh et al. [11], emergency hemodialysis was undertaken for threatening hyperkalemia in 46% of cases, acute lung edema in 20%, poorly tolerated uremic syndrome and anuria were the indications in 16% and 12% respectively, malignant hypercalcemia in 4% and acidosis in 2%.
Hyperkalemia is one of the often-threatening complications associated with CKD and is maintained by metabolic acidosis [16].
The mean creatinine clearance was 5.475 ± 3.368 ml/min/1.73 m2 (Table 3). These results corroborate those of B. Saad et al. [5] in Morocco in 2016, who found a mean of 6.337 ± 0.616 ml/min/1.73 m2. However, they remain lower than those of F. Bernadette et al. [17] in Colmar, France, who found an average of 9.27 ± 2.69 ml/min/1.73 m2. Here, we note an approximation between the figures for the two African countries, which remain lower than those for a region of France. This constitutes biological proof of the much more pronounced delay in initiating hemodialysis in African regions compared with developed countries.
In terms of hemoglobin level, 91.8% of our patients had a level < 10 g/dl, 12.5% had between 5 and 6.9 g/dl (Table 3). In 2011 in Côte d’Ivoire, B. Ouattara et al. [18] found that the mean hemoglobin level in their patients was 7.6 ± 2.2 g/dl. The anemia of CKD patients could be explained by erythropoietin deficiency, the intra-dialytic blood loss, during everyday events (e.g. road accidents), advanced uremia, vitamin B12 and folate deficiency [18]-[20].
In relation to initial kidney disease, chronic glomerulonephritis was the main cause of CKD in our study (45.7%), followed by vascular nephropathy (25.7%), and then diabetic nephropathy (11.4%) (Table 3). These results concur with those of F. G. Sane et al. [21] in Senegal in 2019, who identified chronic glomerulonephritis (45.5%) and nephroangiosclerosis (26.5%) as the main causes of CKD. Undetermined nephropathy was noted in only one of our patients (Table 3). This result is much lower than that of F. G. Sane et al. [21], who reported 12.5% of cases. Our results differ from those of B. Saad et al. [5] in Morocco, who reported in their study that diabetic nephropathy was the most frequent cause of CKD, accounting for 41% of cases, followed by other glomerular and tubular nephropathies. The initial nephropathy was not known in 13% of cases, due to late discovery of CKD. Our results show a clear reduction in the rate of undetermined nephropathy. This is probably due not only to a better knowledge of kidney pathologies by medical staff, but also to an improvement in the means of exploring these pathologies in our country. We must also bear in mind that our sample was small and the study was limited in time, which could constitute a bias that would prevent us from drawing any conclusions.
In the management of end-stage renal disease, the main problem is the late referral of patients to specialist consultations. Outside our case, not only is this situation frequent, with rates varying from one country to another and from one series to another (between 20% and 82%), but it also depends on the local definition of “late referral”. [19]
In our study, only 18.6% of patients were followed up early before hemodialysis, giving a late referral rate of 81.4% (Table 3). This rate of late referral (LR) is higher than that of other European and American series published after 1995 [22] [23], which ranged from 20% to 60%. Our result is also higher than that of M. Raffray et al. [24] in their study on “...the determinants of an emergency start of dialysis” published in 2017, who reported that 30% of patients with end-stage chronic renal failure (ESRD) in France started hemodialysis in an emergency.
This high rate of ESRD in our study can be explained, on the one hand, by the refusal of some patients to accept the diagnosis, thus delaying treatment, and, on the other, by the high level of poverty among our populations in relation to the cost of paraclinical examinations or treatments, or by the lack of functioning dialysis centers in Guinea, both public and private, which are all concentrated in Conakry apart from one small center in the north-east of the country. All these factors led patients to seek treatment only in cases of extreme urgency.
The main indication for hemodialysis in our study was uremic intoxication (74.3%), followed by hyperkalemia (31.4%) (Table 3). The same trend was noted in the study by Diawara et al. [8] with 50.47% for poorly tolerated uremia and 40.19% for threatening hyperkalemia.
These results corroborate data from the literature, for which uremic intolerance, hyperkalemia and acute lung edema are among the main indications for extra-renal purification [19]. The presence of these complications at the start of hemodialysis in our patients could be explained not only by the delay in consultation with the nephrologist, but also by the quality of follow-up, with patients in some cases lost to sight of one another in the pre-terminal or terminal period of the disease, and finally by the remoteness of dialysis centers from the populations of the interior of the country, these functional centers being concentrated in Conakry for the time being.
In our study, 100% of patients were initially managed using a central catheterization vascular approach. Two types of temporary vascular approach were the main way used in 57 cases of late referral patients: femoral and simple jugular catheterization (84.2% and 10.5% respectively). Thus, only 3 early and 3 late referrals were treated by tunneled jugular catheterization, which was the only permanent vascular approach used in our study. For the remaining 10 early referral patients, there were 5 cases of femoral catheterization versus 5 cases of simple jugular catheterization (Table 3). Although some patients with old arteriovenous fistulas (AVFs) continued to undergo hemodialysis at the center, there were no cases of new fistulas during our study period, either because our patients’ free choice was not favorable, or because the quality of the vascular capital was poor, or because there were financial difficulties for the surgery, the costs being borne by the patient. It should be noted that there is no competence for prosthetic catheterization locally. Our results are similar to those of Y. Kane et al. [25] in Senegal published in 2017, for whom 91.7% of patients had had a central venous catheter as their first vascular approach. Similarly, in the work by Abdelaali Behadi et al. [26] in Morocco published the same year, only 105 patients (33%) had received specialized nephrological follow-up, and in almost two-thirds of cases, hemodialysis had been started using a temporary, mainly femoral, central venous catheter.
According to DOQI recommendations [27], the proportion of AVFs used as first vascular access in subjects starting hemodialysis should be 50%. The DOQI’s revised hemodialysis good practice guide of 2000 [28] sets a target of less than 20% temporary central catheters in a hemodialysis unit. With the statistics listed above, it is clear that the majority of our African regions are still a long way from meeting these recommendations.
Indeed, the health policy of most countries in the south of the Sahara is like that of Guinea, which we know more about. It explains the high frequency of late referral of already complicated CKD patients to specialist facilities. As a consequence of these health policies, there’s no state provision for these already poor populations, and the health insurance systems are almost non-functional. This de facto should explain the high proportion of temporary catheterization, particularly femoral catheterization for initiation of dialysis. This type of vascular approach is the easiest, least prone to complications during the surgery, and least costly, making it suitable for emergency dialysis [29] [30].
In the bivariate analysis, with regard to comorbidities, there were more diabetic patients in the late than early group (7 (58.3%) vs 5 (41.7%); p value = 0.02) (Table 3). However, according to the work of M. Hoffmann et al. [31], diabetics are the most at-risk chronic kidney disease patients, followed by a nephrologist, and they should receive optimized specialized care. This was not the case with our diabetic patients, the majority of whom were in the late referral group, having started dialysis without prior long-term follow-up.
Uremic intoxication was more frequent in the late referral group (39 (75.0%); p value = 0.02) (Table 3).
Hypercreatininemia was statistically significant, much higher in the early reference group (198.270 ± 101.230 mg/L; p = 0.020) (Table 3). We found no explanation for this trend towards higher hypercreatininemia in early referrals compared with late referrals.
With the exception of six, all our patients started hemodialysis on a temporary vascular approach of the femoral (p value = 0.001) and simple jugular (p value = 0.013) catheterization type. As shown above, the only permanent vascular approach used in our patients during the study period for initiation of hemodialysis was tunnel catheterization, with three cases of late referral and three cases of early referral (Table 3). There were no cases of arteriovenous fistula (AVF), arteriovenous bypass (AVB) or prosthetic goretex (PTFE) bypass at the start of hemodialysis.
To conclude this part, therefore, we will say this: after bivariate analysis, it was found that there were statistically significant associations between late referral and certain factors such as diabetes comorbidity (p = 0.02), uremic intoxication (p = 0.02), temporary central catheterization (femoral with p = 0.001 and jugular with p = 0.013). Permanent central catheterization or tunneled catheterization was also statistically significant, with identical p values in both groups (p = 0.03) (Table 3).
Finally, the late referral group was significantly more exposed to mortality (4 cases vs. 0 cases) during the first three months of dialysis, although there was no established statistical link between referral mode and mortality (p = 2) (Table 4). M. Raffray et al. [24] noted in a 2017 study in France that emergency start-up of dialysis increased the risk of morbidity and mortality, and reduced patients’ quality of life. As for M. Hoffmann et al. [31], in their work published in 2006, mortality among dialysis patients is said to have “remained high and relatively constant over the last ten years”. Among the many causes put forward, one of the most incriminating factors was late referral to the nephrologist. This late referral of CKD patients was associated with early excess mortality limited to the first three months post-dialysis. It accounted for around 20% of the patients managed on dialysis each year.
Table 4. Impact of early and late follow-up on mortality in CKD patients initiated on hemodialysis.
|
Initiation |
p Value |
Late |
Early |
Favorable |
53 (93%) |
13 (100%) |
0.3 |
Deaths |
4 (7%) |
0 (0%) |
2 |
We should also point out that our study only covered a three-month period. It may be possible to obtain different results (in terms of morbidity and mortality) in the future if the same study, perhaps multicentric, is repeated one day on our premises or elsewhere, with larger numbers and over a longer period.
5. Conclusion
Early nephrological follow-up is of fundamental importance for the smooth initiation of hemodialysis. Late referral of patients to the nephrologist, most often in a highly uremic state, has a serious impact not only on their quality of life, but also on medium and long-term morbidity and mortality. In the future, repeating this study with larger numbers and a sufficiently long study period will enable us to reach a more reliable conclusion.