Outcome of diabetic hyperglycaemic emergencies in a Nigerian cohort
http://www.100md.com
Okoro EO1, Yusuf M1, Salawu HO1,Oyejola
diabetic,hyperglycaemic,emergency;outcome;,Nigeria,OutcomeofdiabetichyperglycaemicemergenciesinaNigeriancohort,INTRODUCTION,METHODS,DataCollection,St
参见附件。
Outcome of diabetic hyperglycaemic emergencies in a Nigerian cohort(pdf)
Correspondence to Dr E.O Okoro, Department of Medicine, University of Ilorin PMB 1515, Ilorin, Kwara State, Nigeria
Tel:+2348055933574
E-mail: eookoro2003@yahoo.co.uk
[Abstract] Objective Outcome measures of patients admitted with acute diabetic hyperglycaemic emergencies in a teaching hospital were observed over a forty month period in order to determine the effect of a case management based on patient's ability to pay. Methods Individuals admitted with diabetic ketoacidosis (DKA) or hyperosmolar non-ketosis (HONK) were managed largely by clinical parameters using only those laboratory tests available and can be paid for by the patient Results Forty of diabetic related admissions out of 120 were for DKA and HONK. Thirty-two were for DKA and 8 for HONK. Subjects with HONK were significantly (P= 0.043) older when compared to individuals with DKA, being 58.10 + 10.4 vs 44.8 + 17 years respectively. There were 22 survivors out of 32 DKA and 6 survivors of those admitted with HONK. Excluding three individuals who discharged themselves against medical opinion, the crude mortality rate for DKA was 22% and 25% for HONK. Two of the nine deaths occurred within 48 hours of hospitalization. Conclusions These figures are troubling but the data also suggests that creativity and prudence in the utilization of available resources can improve outcome.
[Key words] diabetic hyperglycaemic emergency;outcome; Nigeria
INTRODUCTION
Previous reports [1,2] from Nigeria suggest that diabetic hyperglycaemic emergencies were frequently associated with a poor outcome. This has been variously attributed [1,2] to late patient presentation in hospital; inadequate laboratory services and inability of patients to pay for such services even when available.
However, since 2002 as part of efforts [3] to improve quality of diabetic care, we have operated with two cardinal benchmarks; i.e. (i) any individual attending our diabetic clinic for at least six months; and/or (ii) those who can pay the admission fees and die from either diabetic ketoacidosis (DKA) or hyperosmolar non-ketotic (HONK) irrespective of other contributing factors,was regarded as a failure of the care we give.
These in-house standards were adopted to encourage the use of the opportunity offered by clinic visits for regular patient education and also to encourage creativity in the utilization of scarce resources[4] given the endemic nature of poverty and other constraints in our operating environment. This report therefore presents our observation in the forty months under consideration i. e. April to August 2005.
METHODS
Setting Ours is a university teaching hospital based facility that provides health services on a fee-for-service basis [5].Consultant Physicians and trainee specialists in the department are grouped into functional units called firms/divisions with two of the three Consultants in one of the five firms/divisions rotating the duties between them weekly to provide diabetic care services to the hospital on a honorary basis. The firm also keeps a logbook that summarizes admissions, diagnosis, outcome and other information.
Data Collection
Information on patients admitted for diabetes related conditions from May 1，2002 to August 30, 2005 was extracted from their case notes and complimented with those in the logbook. For the purpose of this study, DKA was diagnosed in the presence of elevated blood glucose ≥ 13 mmol, ketonuria of at least (++) by dipstick and clinical dehydration [6]. HONK was diagnosed in the presence of clinical dehydration, scanty (+) or absent ketonuria by urinalysis and an effective serum osmolality of ≥ 340 mOsm/ L, calculated using the formula: 2[Na+ + K+] + Glucose [7, 8, 9]. Where HONK was suspected on clinical grounds but serum electrolytes and glucose could not be measured because of problems in the laboratory e.g. interruption of electricity supply or equipment breakdown, such cases were regarded as poorly controlled diabetes and managed as such.
Statistics
Mean comparison of two groups was carried out using t-tests and comparison of two populations was done using binomial test. Differences were considered significant if P was < 0.05 except as otherwise indicated. Statistical software package SPSS version 11.0 Windows were used.
Clinical Case Management
Suspected cases were admitted into the regular ward after confirmation with dipstick urinalysis and RBG( Random blood glucose) often using a hand-held glucometer that has become available since 2004 for pin-prick capillary blood glucose estimation at a bedside cost of N200 (1.4 US Dollars) per test. Blood samples were also taken for baseline measurements of serum Na+, K+, RBG and occasionally urea when indicated and available. Soon after, one litre of normal saline was given intravenously (i.v.) in the first hour, then 250~500 ml/h thereafter as needed and as determined by the clinical severity of dehydration e.g. the presence or otherwise of orthostatic hypotension, mucosal dryness, loss of skin elasticity, oliguria etc. and titrated to achieve a urine output of 1~3 ml/kg/h.
An initial bolus of 5~10 units of soluble insulin was also given i.v. and then added to the saline infusion at an initial dose of 0.1 unit/kg/h and titrated until the urine is ketone free on two consecutive occasions. For this latter, double voided urine was used for testing. Normal saline as changed to glucose-containing fluids e.g. 5% dextrose in 0.45 saline or 10% dextrose if blood glucose dropped to 11~13 mmol/L and ketonuria persisted. As soon as urine output was adequate as defined above, 20 mEq of KCl was added to each litre of infusion and neither food nor drink was allowed during this period. In some cases and as indicated by the presence of gastric succussion splash, nasogastric tube was passed to empty the stomach.
Patients were monitored for the presence of basal crepitations, elevated JVP and increasing respiratory rate. The rate of infusion was adjusted accordingly to prevent fluid overload. Fluid intake and output were also charted by nurses and at times assisted by the patient's relations as a way of monitoring fluid utilization. Febrile patients were treated for acute malaria and antibiotics were given empirically only if they were suggestive localizing signs e.g. abnormal chest findings, dysuria or suprapubic tenderness.
Follow-up tests and their frequency were generally guided by the patient's ability to pay for them and their availability at the time of hospitalization. For those who could afford the cost and the service was available, hourly RBG was done and when blood glucose dropped to 11～13 mmol/L or after about 3～4 hours and an improvement in clinical status (for those with financial constraint) was noticed, the fluid was usually changed to 5% dextrose in 0.45% saline and the insulin-K+ infusion continued until the urine is ketone free on two consecutive occasions. Specifically, the timing of the follow-up urinalysis was usually guided by improvement in patient's hydration status and general well being. As soon as the patient's hydration status was adequate, graded oral fluids were introduced and intraveneus infusion continued until there was a request for food. Once food was tolerated, the infusion was discontinued after half an hour of subcutaneous insulin and oral KCl supplements such as Slow-K had been introduced. Oral hypoglycaemic agents and/or lente insulin were also usually started at this stage based on body mass index or previous medication history [3,5].Treatment of HONK generally followed the same pattern except that fluid administration was often less intensive and insulin was commenced at 0.05 units/kg/h. A member of the medical team was often designated to closely monitor the patient's progress and the strip for urinalysis was often split longitudinally into two in order to obtain two tests for the price of one.
RESULTS
In the period studied, 1,124 individuals were hospitalized of whom 120(10.6%) were diabetes related. Specifically, 40 of these i.e. 33.3% were for acute hyperglycaemic syndromes/ emergencies, of DKA [32] and HONK [8]. The demographic and other selected variables of the patients are set out in Table 1. Essentially, 8 of these individuals with type 1 diabetes were younger (P<0.001) with a mean age of 20.38 ± 4.8 years compared with those with type 2 diabetes who had a mean age of 54.13 ± 10.53 years. Duration of disease was similar (P=0.64) in both groups, being 7.5 ± 4.3 and 9.03 ± 8.86 years for type 1 and type 2 diabetes respectively. In this cohort, 18 were attending our diabetic clinic for six months or longer prior to their admission (see INTRODUCTION section). The data summarized in Table 2 showed that DKA was the most common (86%) acute hyperglycaemic emergency seen in our facility with a crude case fatality rate of 21.9% and most of the deaths occurred after two days in hospital (Tables 1,2). For HONK, the figure was 25%. Of the 28 survivors,six of them were young people with type 1 disease and DKA (Tables 1,3). The only death in this group occurred on the first day of admission (Table 1). Three individuals (7.5% of the cohort) discharged themselves against medical advice (DAMA) after a mean hospital stay of 4 ±1 days (see Table 1). This is in contrast to a mean hospital stay of 9.89 ± 7.9 for survivors as against 6.67 ± 9.64 days (P=0.32) for those who died.
Table 1 Selected Variables of Individual Patients
Continued
Note:M-male, F-female;DKA-diabetic ketoacidosis;HONK-hyperosmolar non-ketosis;DAMA-discharged against medical advice
Table 2 Outcome of Diabetic Hyperglycaemic Syndromes by Metabolic Type
Note:ΨUsing t-test,*significant difference
Table 3 Outcome by Diabetes Type
Note:ΨUsing t-test,*significant difference
DISCUSSION
The case fatalities of about 22% for DKA and 25% for HONK herein observed are troubling when compared to rates of less than 5% and 17.5% for DKA and HONK respectively generally reported in many countries with advanced health care systems [6, 7, 10]. Even so, 45% of the individuals in this cohort had attended our diabetic clinic at some point prior to their hospitalization, thus signaling a possible failure of our educational and preventive programmes for individuals with diabetes [7, 11]. Of note, only two of the nine deaths occurred within 48 hours of admission (Table 2) and this could suggest that the processes of care might also have played some role in the pattern of mortality observed.
However, it is important to interpret these findings in the context of a number of issues. First, none of the cases herein reported involved ICU management; the use of Arterial Blood Gas (ABG) analysis, CVP line; nor insulin pump often advocated as part of standard practice [6, 8, 9]. Second, the frequency of laboratory tests such as hourly blood electrolytes, glucose and urinalysis so recommended in standard clinical protocols in USA, Europe and Middle East [6,7, 11] was abridged, as most of the test items were either not available or could not be paid for by the patients at the frequency recommended in such protocols. Consequently, for many patients, therapy was guided predominantly by clinical parameters and accumulated experience [13, 14]. Of note, effective osmolality which truly reflects tonicity was adopted rather than osmolarity which is its concentration and incooperates urea which diffuses freely across cell membrane and adds little to osmolality [9, 10] but increases cost of care.
Third, frequent blood glucose monitoring and urine testing with adjustment of medications at home which make DKA and HONK less likely to occur in those living with diabetes was not a strategy that is financially feasible for most of our patients [3, 7, 11] because of cost and widespread poverty.
Fourth, comparative historic data from many African countries at similar stage of development suggest a case fatality of between 33% and 50% for DKA and HONK[12, 13, 15]. Even within Nigeria, regional variations appear to exist in the outcome of patients with these conditions admitted in similar hospitals. For example, a case fatality rate of about 54.5% was reported for DKA and HONK in the Niger-Delta, with many of those in that cohort presenting for the very first time[1]. Similarly, a case fatality rate of 50% for HONK was observed [16] in Ibadan, which is about 200km south west of this centre. More recently, Physicians in another Teaching Hospital 350km south of our location reported [17] that 29.3% of all acute diabetic deaths resulted from DKA and HONK. Seen in this light, the mortality observed in this cohort does not appear worse than those previously reported from similar centres. If anything, our data suggest some improvement particularly as most of the known patients were also chronic clinic defaulters while other (the majority) were presenting for the very first time. Clearly, there are severe resource constraints in our practice environment but this figure can be better as the present experience also suggests that creativity and prudence are imperative in making optimal use of available resources in enhancing performance and improving quality of diabetic care. This experience and the present findings will now be used in developing a management protocol for DKA and HONK that is relevant to our operating environment and other resource poor settings and serve as baseline for evaluating subsequent interventions.
ACKNOWLEDGEMENT
We thank our Colleagues who were involved in management of some of these cases. The help of Mr. Akande, Oladapo Olaniyi in preparing the script is gratefully acknowledged.
REFERENCES
1. Dagogo Jack S. Survey of Diabetic-in-patient mortality in Port-Harcourt Nigeria. Orient J Medicine, 1991,3(1):37-41.
2. Ndububa DA, Erharbor GE. Diabetes mortalities in Ilesa, Nigeria: a retrospective study. Centr Afr J Med, 1994,40(10): 291-292.
3. Okoro EO, Adejumo AO, Oyejola BA. Quality of diabetic care in Nigeria: report of a self audit. J Diabetes Complications,2002,16(2): 159-164.
4. Muula A. preventing diabetes associated morbidity and mortality in resource poor communities. Diabetes Int,2000,10(2): 47-49.
5. Okoro EO, Oyejola BA. Inadequate control of blood pressure in Nigerians with diabetes. Ethnicity & Disease, 2004, 14: 82-86.
6. Khan SA, Khan LA. Presentation and outcome of diabetic ketoacidosis in Nigerians.Saudi Arabia Diabetes International,2000,10(2):50-51.
7. Magee MF, Bhatt BA. Management of decompensated Diabetes. Critical care clinics, 2001, 17(1):75-106.
8. Matz R. Management of the hyperosmolar hyperglycaemic syndrome. Am Fam Phys, 1999, 60:1468-1476.
9. Umpierrez GE, Khajari M, Kitabichi AE.Diabetic Ketoacidosis and hyperglycaemic hyperosmolar non-ketotic syndrome.Am J med Sci,1996, 311: 225-233.
10. Basu A, Close CF, Jenkins D,et al.Persisting mortality in diabetic ketoacidosis. Diabetic Medicine, 1993,10(8): 782-784.
11. Mortality indiabetic ketoacidosis.http://www.emedicine.com/ (accessed on 7/1/2006).
12. McLarty DE, Pollit C, Sivai ABM. Diabetes in Africa. Diabetic Medicine, 1990, 7: 670-684.
13. Thomas SE. A review of diabetic hyperglycaemic comas at Kenyatta National Hospital. 1978-80, East Afr Med J,1980, 57: 877-882.
14. Haile S. Management of diabetic ketoacidosis with lente insulin. Diabetes International 2000,10(2): 61.
15. Adubofour KOM, Ofei F, Mensah-Adubofour J, et al.Diabetes in Ghana: a morbidity and mortality analysis. Int Diabetes Digest,1993,4(3): 90-93.
16. Talabi OA. Is hypokalaemia the killer in diabetic hyperosmolar states during treatment? Diabetes International,2000, 10(2): 60-61.
17. Ogbera AO, Ohwovoriole AE, Soyebi O. Case fatality among diabetic-in-patients. J Clinical Sciences, 2002,2(1-2): 18-21.
1 Department of Medicine,University of Ilorin Teaching Hospital, Ilorin, Nigeria
2 Department of Statistics, University of Ilorin, Nigeria
(Editor Emilia)
Outcome of diabetic hyperglycaemic emergencies in a Nigerian cohort(pdf)
Correspondence to Dr E.O Okoro, Department of Medicine, University of Ilorin PMB 1515, Ilorin, Kwara State, Nigeria
Tel:+2348055933574
E-mail: eookoro2003@yahoo.co.uk
[Abstract] Objective Outcome measures of patients admitted with acute diabetic hyperglycaemic emergencies in a teaching hospital were observed over a forty month period in order to determine the effect of a case management based on patient's ability to pay. Methods Individuals admitted with diabetic ketoacidosis (DKA) or hyperosmolar non-ketosis (HONK) were managed largely by clinical parameters using only those laboratory tests available and can be paid for by the patient Results Forty of diabetic related admissions out of 120 were for DKA and HONK. Thirty-two were for DKA and 8 for HONK. Subjects with HONK were significantly (P= 0.043) older when compared to individuals with DKA, being 58.10 + 10.4 vs 44.8 + 17 years respectively. There were 22 survivors out of 32 DKA and 6 survivors of those admitted with HONK. Excluding three individuals who discharged themselves against medical opinion, the crude mortality rate for DKA was 22% and 25% for HONK. Two of the nine deaths occurred within 48 hours of hospitalization. Conclusions These figures are troubling but the data also suggests that creativity and prudence in the utilization of available resources can improve outcome.
[Key words] diabetic hyperglycaemic emergency;outcome; Nigeria
INTRODUCTION
Previous reports [1,2] from Nigeria suggest that diabetic hyperglycaemic emergencies were frequently associated with a poor outcome. This has been variously attributed [1,2] to late patient presentation in hospital; inadequate laboratory services and inability of patients to pay for such services even when available.
However, since 2002 as part of efforts [3] to improve quality of diabetic care, we have operated with two cardinal benchmarks; i.e. (i) any individual attending our diabetic clinic for at least six months; and/or (ii) those who can pay the admission fees and die from either diabetic ketoacidosis (DKA) or hyperosmolar non-ketotic (HONK) irrespective of other contributing factors,was regarded as a failure of the care we give.
These in-house standards were adopted to encourage the use of the opportunity offered by clinic visits for regular patient education and also to encourage creativity in the utilization of scarce resources[4] given the endemic nature of poverty and other constraints in our operating environment. This report therefore presents our observation in the forty months under consideration i. e. April to August 2005.
METHODS
Setting Ours is a university teaching hospital based facility that provides health services on a fee-for-service basis [5].Consultant Physicians and trainee specialists in the department are grouped into functional units called firms/divisions with two of the three Consultants in one of the five firms/divisions rotating the duties between them weekly to provide diabetic care services to the hospital on a honorary basis. The firm also keeps a logbook that summarizes admissions, diagnosis, outcome and other information.
Data Collection
Information on patients admitted for diabetes related conditions from May 1，2002 to August 30, 2005 was extracted from their case notes and complimented with those in the logbook. For the purpose of this study, DKA was diagnosed in the presence of elevated blood glucose ≥ 13 mmol, ketonuria of at least (++) by dipstick and clinical dehydration [6]. HONK was diagnosed in the presence of clinical dehydration, scanty (+) or absent ketonuria by urinalysis and an effective serum osmolality of ≥ 340 mOsm/ L, calculated using the formula: 2[Na+ + K+] + Glucose [7, 8, 9]. Where HONK was suspected on clinical grounds but serum electrolytes and glucose could not be measured because of problems in the laboratory e.g. interruption of electricity supply or equipment breakdown, such cases were regarded as poorly controlled diabetes and managed as such.
Statistics
Mean comparison of two groups was carried out using t-tests and comparison of two populations was done using binomial test. Differences were considered significant if P was < 0.05 except as otherwise indicated. Statistical software package SPSS version 11.0 Windows were used.
Clinical Case Management
Suspected cases were admitted into the regular ward after confirmation with dipstick urinalysis and RBG( Random blood glucose) often using a hand-held glucometer that has become available since 2004 for pin-prick capillary blood glucose estimation at a bedside cost of N200 (1.4 US Dollars) per test. Blood samples were also taken for baseline measurements of serum Na+, K+, RBG and occasionally urea when indicated and available. Soon after, one litre of normal saline was given intravenously (i.v.) in the first hour, then 250~500 ml/h thereafter as needed and as determined by the clinical severity of dehydration e.g. the presence or otherwise of orthostatic hypotension, mucosal dryness, loss of skin elasticity, oliguria etc. and titrated to achieve a urine output of 1~3 ml/kg/h.
An initial bolus of 5~10 units of soluble insulin was also given i.v. and then added to the saline infusion at an initial dose of 0.1 unit/kg/h and titrated until the urine is ketone free on two consecutive occasions. For this latter, double voided urine was used for testing. Normal saline as changed to glucose-containing fluids e.g. 5% dextrose in 0.45 saline or 10% dextrose if blood glucose dropped to 11~13 mmol/L and ketonuria persisted. As soon as urine output was adequate as defined above, 20 mEq of KCl was added to each litre of infusion and neither food nor drink was allowed during this period. In some cases and as indicated by the presence of gastric succussion splash, nasogastric tube was passed to empty the stomach.
Patients were monitored for the presence of basal crepitations, elevated JVP and increasing respiratory rate. The rate of infusion was adjusted accordingly to prevent fluid overload. Fluid intake and output were also charted by nurses and at times assisted by the patient's relations as a way of monitoring fluid utilization. Febrile patients were treated for acute malaria and antibiotics were given empirically only if they were suggestive localizing signs e.g. abnormal chest findings, dysuria or suprapubic tenderness.
Follow-up tests and their frequency were generally guided by the patient's ability to pay for them and their availability at the time of hospitalization. For those who could afford the cost and the service was available, hourly RBG was done and when blood glucose dropped to 11～13 mmol/L or after about 3～4 hours and an improvement in clinical status (for those with financial constraint) was noticed, the fluid was usually changed to 5% dextrose in 0.45% saline and the insulin-K+ infusion continued until the urine is ketone free on two consecutive occasions. Specifically, the timing of the follow-up urinalysis was usually guided by improvement in patient's hydration status and general well being. As soon as the patient's hydration status was adequate, graded oral fluids were introduced and intraveneus infusion continued until there was a request for food. Once food was tolerated, the infusion was discontinued after half an hour of subcutaneous insulin and oral KCl supplements such as Slow-K had been introduced. Oral hypoglycaemic agents and/or lente insulin were also usually started at this stage based on body mass index or previous medication history [3,5].Treatment of HONK generally followed the same pattern except that fluid administration was often less intensive and insulin was commenced at 0.05 units/kg/h. A member of the medical team was often designated to closely monitor the patient's progress and the strip for urinalysis was often split longitudinally into two in order to obtain two tests for the price of one.
RESULTS
In the period studied, 1,124 individuals were hospitalized of whom 120(10.6%) were diabetes related. Specifically, 40 of these i.e. 33.3% were for acute hyperglycaemic syndromes/ emergencies, of DKA [32] and HONK [8]. The demographic and other selected variables of the patients are set out in Table 1. Essentially, 8 of these individuals with type 1 diabetes were younger (P<0.001) with a mean age of 20.38 ± 4.8 years compared with those with type 2 diabetes who had a mean age of 54.13 ± 10.53 years. Duration of disease was similar (P=0.64) in both groups, being 7.5 ± 4.3 and 9.03 ± 8.86 years for type 1 and type 2 diabetes respectively. In this cohort, 18 were attending our diabetic clinic for six months or longer prior to their admission (see INTRODUCTION section). The data summarized in Table 2 showed that DKA was the most common (86%) acute hyperglycaemic emergency seen in our facility with a crude case fatality rate of 21.9% and most of the deaths occurred after two days in hospital (Tables 1,2). For HONK, the figure was 25%. Of the 28 survivors,six of them were young people with type 1 disease and DKA (Tables 1,3). The only death in this group occurred on the first day of admission (Table 1). Three individuals (7.5% of the cohort) discharged themselves against medical advice (DAMA) after a mean hospital stay of 4 ±1 days (see Table 1). This is in contrast to a mean hospital stay of 9.89 ± 7.9 for survivors as against 6.67 ± 9.64 days (P=0.32) for those who died.
Table 1 Selected Variables of Individual Patients
Continued
Note:M-male, F-female;DKA-diabetic ketoacidosis;HONK-hyperosmolar non-ketosis;DAMA-discharged against medical advice
Table 2 Outcome of Diabetic Hyperglycaemic Syndromes by Metabolic Type
Note:ΨUsing t-test,*significant difference
Table 3 Outcome by Diabetes Type
Note:ΨUsing t-test,*significant difference
DISCUSSION
The case fatalities of about 22% for DKA and 25% for HONK herein observed are troubling when compared to rates of less than 5% and 17.5% for DKA and HONK respectively generally reported in many countries with advanced health care systems [6, 7, 10]. Even so, 45% of the individuals in this cohort had attended our diabetic clinic at some point prior to their hospitalization, thus signaling a possible failure of our educational and preventive programmes for individuals with diabetes [7, 11]. Of note, only two of the nine deaths occurred within 48 hours of admission (Table 2) and this could suggest that the processes of care might also have played some role in the pattern of mortality observed.
However, it is important to interpret these findings in the context of a number of issues. First, none of the cases herein reported involved ICU management; the use of Arterial Blood Gas (ABG) analysis, CVP line; nor insulin pump often advocated as part of standard practice [6, 8, 9]. Second, the frequency of laboratory tests such as hourly blood electrolytes, glucose and urinalysis so recommended in standard clinical protocols in USA, Europe and Middle East [6,7, 11] was abridged, as most of the test items were either not available or could not be paid for by the patients at the frequency recommended in such protocols. Consequently, for many patients, therapy was guided predominantly by clinical parameters and accumulated experience [13, 14]. Of note, effective osmolality which truly reflects tonicity was adopted rather than osmolarity which is its concentration and incooperates urea which diffuses freely across cell membrane and adds little to osmolality [9, 10] but increases cost of care.
Third, frequent blood glucose monitoring and urine testing with adjustment of medications at home which make DKA and HONK less likely to occur in those living with diabetes was not a strategy that is financially feasible for most of our patients [3, 7, 11] because of cost and widespread poverty.
Fourth, comparative historic data from many African countries at similar stage of development suggest a case fatality of between 33% and 50% for DKA and HONK[12, 13, 15]. Even within Nigeria, regional variations appear to exist in the outcome of patients with these conditions admitted in similar hospitals. For example, a case fatality rate of about 54.5% was reported for DKA and HONK in the Niger-Delta, with many of those in that cohort presenting for the very first time[1]. Similarly, a case fatality rate of 50% for HONK was observed [16] in Ibadan, which is about 200km south west of this centre. More recently, Physicians in another Teaching Hospital 350km south of our location reported [17] that 29.3% of all acute diabetic deaths resulted from DKA and HONK. Seen in this light, the mortality observed in this cohort does not appear worse than those previously reported from similar centres. If anything, our data suggest some improvement particularly as most of the known patients were also chronic clinic defaulters while other (the majority) were presenting for the very first time. Clearly, there are severe resource constraints in our practice environment but this figure can be better as the present experience also suggests that creativity and prudence are imperative in making optimal use of available resources in enhancing performance and improving quality of diabetic care. This experience and the present findings will now be used in developing a management protocol for DKA and HONK that is relevant to our operating environment and other resource poor settings and serve as baseline for evaluating subsequent interventions.
ACKNOWLEDGEMENT
We thank our Colleagues who were involved in management of some of these cases. The help of Mr. Akande, Oladapo Olaniyi in preparing the script is gratefully acknowledged.
REFERENCES
1. Dagogo Jack S. Survey of Diabetic-in-patient mortality in Port-Harcourt Nigeria. Orient J Medicine, 1991,3(1):37-41.
2. Ndububa DA, Erharbor GE. Diabetes mortalities in Ilesa, Nigeria: a retrospective study. Centr Afr J Med, 1994,40(10): 291-292.
3. Okoro EO, Adejumo AO, Oyejola BA. Quality of diabetic care in Nigeria: report of a self audit. J Diabetes Complications,2002,16(2): 159-164.
4. Muula A. preventing diabetes associated morbidity and mortality in resource poor communities. Diabetes Int,2000,10(2): 47-49.
5. Okoro EO, Oyejola BA. Inadequate control of blood pressure in Nigerians with diabetes. Ethnicity & Disease, 2004, 14: 82-86.
6. Khan SA, Khan LA. Presentation and outcome of diabetic ketoacidosis in Nigerians.Saudi Arabia Diabetes International,2000,10(2):50-51.
7. Magee MF, Bhatt BA. Management of decompensated Diabetes. Critical care clinics, 2001, 17(1):75-106.
8. Matz R. Management of the hyperosmolar hyperglycaemic syndrome. Am Fam Phys, 1999, 60:1468-1476.
9. Umpierrez GE, Khajari M, Kitabichi AE.Diabetic Ketoacidosis and hyperglycaemic hyperosmolar non-ketotic syndrome.Am J med Sci,1996, 311: 225-233.
10. Basu A, Close CF, Jenkins D,et al.Persisting mortality in diabetic ketoacidosis. Diabetic Medicine, 1993,10(8): 782-784.
11. Mortality indiabetic ketoacidosis.http://www.emedicine.com/ (accessed on 7/1/2006).
12. McLarty DE, Pollit C, Sivai ABM. Diabetes in Africa. Diabetic Medicine, 1990, 7: 670-684.
13. Thomas SE. A review of diabetic hyperglycaemic comas at Kenyatta National Hospital. 1978-80, East Afr Med J,1980, 57: 877-882.
14. Haile S. Management of diabetic ketoacidosis with lente insulin. Diabetes International 2000,10(2): 61.
15. Adubofour KOM, Ofei F, Mensah-Adubofour J, et al.Diabetes in Ghana: a morbidity and mortality analysis. Int Diabetes Digest,1993,4(3): 90-93.
16. Talabi OA. Is hypokalaemia the killer in diabetic hyperosmolar states during treatment? Diabetes International,2000, 10(2): 60-61.
17. Ogbera AO, Ohwovoriole AE, Soyebi O. Case fatality among diabetic-in-patients. J Clinical Sciences, 2002,2(1-2): 18-21.
1 Department of Medicine,University of Ilorin Teaching Hospital, Ilorin, Nigeria
2 Department of Statistics, University of Ilorin, Nigeria
(Editor Emilia)
您现在查看是摘要介绍页，详见PDF附件。