top of page

Thromboembolic and bleeding complications in patients with prosthetic heart valves at the Kenyatta National Hospital

Hungu EW1 MBChB, MMed (Surg),. Ogendo SWO2 MBChB, MMed (Surg), FCS (ECSA), 1. Lodwar District Hospital, 2. Department of Surgery, University of Nairobi. Corresponding surgeon: Eric Hungu. E-mail:


Background: Despite constant monitoring of anticoagulation in pros-thetic valve patients, haematologic complications occur. This study documented the occurrence of such complications and associated risk factors at the Kenyatta National Hospital (KNH).

Study design: Observational study reviewing 142 patients, 39 pro-spective and 103 retrospective.

Outcome measures: International Normalised Ratio (INR); presence of signs and symptoms of haematological complications.

Results: Forty four (31%) patients presented with bleeding tenden-cies, 28 grade I and 4 grade III. The most common thromboembolic complication was headache in 33 (23.2%) patients.


Mean duration of anticoagulation for patients with complications was 82.9 months (± 64), compared to 60.8 months (± 43.8) in those without.

Nine patients were non-compliant in taking medications, haematologic complications presenting in 8 of them.


Conclusion: A positive association was established between hema-tologic complications and INR levels, duration of anticoagulation therapy, non-compliance in taking of medications, and increased period between clinic visits.


Of these only the duration of anticoagulation was an independent predictor for haematological complications.



Since the first valve replacement, and many modifica-tions since in valves design and surgical techniques, the recurrent problems of anticoagulant related complica-tions continue . Complications are not without risk and can be fatal. This study documented the occurrence of haematologic complications in prosthetic valve patients at the KNH, and risk factors to their occurrence.


Patients and Methods


The study examined prosthetic heart valve patients on outpatient follow up at the KNH. The prospective arm of the study covered 16th July 2010 to 14th January 2011, the retrospective arm 10 years prior to July 21st 2010. Bioprosthetic valves were not included in the study.


Outcome measures were INR results during clinic visits, and development of complications. Complications were signs and symptoms of bleeding tendencies, neurologi-cal deficits suggestive of thromboembolism for instance fainting episodes, headaches, and dizziness (Table 1). Target INR for these patients was 2.5 – 4.5.


Data collected was analysed using STATA version 11 for relationship using bivariate analysis and multivariate analysis for independent predictors of haematological complications. A p value of less than 0.05 was consid-ered significant.



A hundred and forty two patients were enrolled into the study, 39 prospectively and 103 retrospective. Sixty one percent of patients were female. Thirty three percent of the patients had at least a basic education and 21% in formal employment (table 2).


Most common surgery was mitral valve replacement and an almost equal number of aortic and double valve re-placement surgeries (table 3).


More than half the patients returned to the outpatient clinic within 3 months of their previous visit.


Mean INR level for study population was 2.3 ± 0.7. In the prospective arm of the study, 13 patients reported noncompliance or interruption in taking medication for various reasons including dental procedures, during menses, elevated INR and pregnancy. A total of 78 (54.9%) patients developed haematologi-cal complications, of these 20 had both thrombotic and haemorrhagic complications.


Click to view table 1


Majority of patients presented with features of throm-botic complications. Of these the main symptom was headache (figure 1) in 33 (23.2%) patients. Forty four patients (31%) presented with bleeding ten-dencies, out of who twenty eight had grade I and eleven grade III bleeding. Of the patients developing haematological complica-tions there was significant association with the INR lev-els (p = 0.017). More women than men developed com-plications though not statistically significant (p 0.38). Of patients with haematologic complications 54% were educated to secondary level. Mean duration of anticoag-ulation for patients developing complications was 82.9 months (± 64 months), compared to 60.8 months (± 43.8 months) for those without. Of 9 patients non-compliant with medications (not taking), 8 developed haematologic complications. No patient interrupting medication (compliant, but only interrupted from time to time) reported any complica-tions (table 5).



Control of INR remains a challenge in developing coun-tries. Without adequate control haematologic complica-tions are inevitable. This study recorded a complication rate of 54.9% and noncompliance rate of 9%, high by any standards. An association between poor compliance and occurrence of complications is in keeping with oth-er studies. However, patients on short term interruption of medication were not associated with increased risk.


Bivariate analysis shows clinic visit interval as a weak association with development of complications (p =0.052). As admitted by many patients, financial constraints make regular clinic attendance difficult. Cou-pled with this not all patients adhere to their medica-tions as required making compliance and interruption of medication a major problem. We suspect some pa-tients only commence warfarin a week before clinic due date giving INR results suggesting faithful compliance to medication. This behaviour is intended to maintain good doctor patient relationship .


The increased occurrence of complications is likely to be due therefore to a combination of noncompliance and associated to INR fluctuations. Previous studies here have shown a poor ability to maintain adequate anti-coagulation with only a 6.9% of patients maintaining adequate control. Fluctuations of INR were associated with complications and our findings are consistent with similar results.


The duration of anticoagulation therapy is an indepen-dent predictor of haematologic complications. Longer duration of therapy positively associated with develop-ment of complications. A plausible explanation being that with risk of complication being present, the longer the duration the more likely a cumulative increase in the number of complications will occur. A local study showed that adequate anticoagulation was maintained for only 18% of follow up time8.


The level of patient education or socioeconomic status was not a risk factor for development of complications. This is in contrast to a South African study, which dem-onstrated a direct association with occurrence of haema-tologic complications. Generally education and age have a positive association with INR control. Increasing age has been shown to have a positive correlation with INR control. Having mixed patients of generally younger ages and education it is likely this relationship is lost in our analysis.

Click to view figure 1


Click to view table 2

The type of surgery performed was not a risk factor for the occurrence of complications, 54.7% of patients with complications having undergone mitral valve replace-ment surgery. However, this could be attributed to mi-tral valve replacement being the most common surgery. A South African study showed a higher incidence of haematologic complications in patients who had un-dergone double valve replacement than in patients who had single valve surgery.



Our valve patients on warfarin follow up experience a high complication rate that will need to be addressed. Development of complications is associated INR con-


Click to view table 3

trol, compliance to medication, interruption of medi-cation and duration of anticoagulation. Of all these variables only duration of anticoagulation is an inde-pendent predictor of haematologic complications.



Identified risk factors can be remedied by ensuring pa-tients are fully educated on the gravity of complications associated with non-compliance. One option would be stricter patient follow up. However, this places great financial burden on the patients, most of who have to travel very long distances for follow up. Estimations of INR are only possible in Nairobi. For this reason self-management of anticoagulation practiced in the West may not be a solution for us either .


Viable alternative would be the establishment of other laboratories outside Nairobi where patients can have their INR monitored. Establishment of other cardiothoracic centres distributed throughout the country would be the eventual solution.


Click to view table 4

Click to view table 5




  1. Harken DE, Soroff HS, Taylor WJ, et al. Partial and complete prostheses in aortic insufficiency. J Thorac Cardiovasc Surg.. 1960; 40: 744-762.

  2. Stein PD, Alpert JS, Copeland J, et al. Antithrombotic therapy in patients with mechanical and biological prosthetic heart valves. Chest. 1992;102(suppl): 445S-454S

  3. Levine MN, Hirsh J, Landefeld S, et al. Hemorrhagic com-plications of anticoagulant therapy. Chest. 1992;102(suppl): 353S-363S.

  4. Heinrich Kortke, Reiner Korfer. International Normalised ra-tio After Heart Valve Replacement Surgery: Is An Early Start Advantageous? Ann Thorac Surg 2001;72: 44-48

  5. Buchanan-Leel B, Levetan B.N, Lombard C.J, et al. Fixed-dose versus adjusted dose warfarin inpatients with pros-thetic heart valves in a peri-urban impoverished population. J Heart Valve Dis 2002;11 (4):583-592.

  6. Lengyel M, Fuster V, Keltai M, et al. Guidelines for manage-ment of left-sided Prosthetic Valve Thrombosis: A Role for Thrombolytic Therapy. J Am Coll Cardiol 1997;30:1521– 1526.

  7. Lars Osterberg, Terrence Blaschke. Adherence to Medication. N Engl J Med 2005; 353:487-497

  8. Ogendo S.W.O. Pattern of anticoagulation control after heart valve surgery at the Kenyatta National Hospital, Nairobi. East Afr Med J 2000 Jul;77 (7):354-358.

  9. Edmunds L.H. Thrombotic and bleeding complications of prosthetic heart valves. Ann Thorac Surg 1987; 44: 430-445

  10. Schapkaitz E, Jacobson B.F, Becker P, et al. Thromboembolic and bleeding complications in patients with mechanical valve replacements-a prospective observational study. SAMJ 2006 Aug; 96 (8):710-713.

  11. Syed S. Hasan, R. Shamala, Imran A. Syed, et al. Factors Affecting Warfarin-Related Knowledge and INR Control of Patients Attending Physician- and Pharmacist- Managed Anticoagulation Clinics. Journal of Pharmacy Practice 2011;24(5) 485-493

  12. Fitzmaurice D.A, Machin S.J. British Society of Haematology Task Force for Haemostasis Thrombosis. Recommendations for Patients Undertaking Self-Management of Oral Antico-agulation. BMJ 2001;323:985–989.

Read Only >>>
bottom of page