Fungal endocarditis after cardiac valve replacementSuman G Kinare, AP Chaukar, SR Panday, GB Parulkar
Department of Cardiovascular Pathology and Surgery, Seth G.S. Medical College and K.E.M. Hospital, Parel, Bombay-400012, India
Correspondence Address: Source of Support: None, Conflict of Interest: None PMID: 722613
Source of Support: None, Conflict of Interest: None
Fungal endocarditis developed in 15 cases out of 205 in whom cardiac valves were replaced during the years 1969-75. Bacterial and fungal infections co-existed in 3 cases. Pure bacterial endocarditis was seen only in one case during this period. Fungal endocarditis was not observed following any other type of open or closed heart surgery. Clinicopathologic features of these cases are briefly described. Paucity of clinical signs and symptoms make early diagnosis difficult. The diagnosis of fungal infection was essentially histological. It was corroborated by culture in 3 cases. In 6 cases candida was responsible while in others either Aspergillus or Penicillium was responsible. In 11 out of 15 cases, endocarditis supervened in the early post-operative period stressing the need for vigilance.
Mycotic endocarditis has now become one of the formidable complications of open heart surgery. It is particularly common following valve replacement. ,,,,, The administration of broad spectrum antibiotics covering pre and post-operative period has succeeded in controlling bacterial infection to a great extent but unfortunately it helps the establishment of fungal infection. , Treatment of systemic mycosis still unsatisfactory and the mortality is very high . ,, In our experience this complication is proving to be major stumbling block in cardiac valve replacement a present. Early diagnosis of fungal endocarditis would improve the chances of survival. From this point of view the clinicopathological pattern presented by cases under our observation was reviewed.
Fifteen cases of post-operative fungal endocarditis which ended fatally during the years 1969-75 form the basis of this study. Autopsy examination was carried out in all the cases. In addition to the detailed morphological study, the vegetations were cultured for bacteria in all the cases and for fungi in 6 cases. Paraffin sections were prepared from all the organs and stained routinely with haematoxylin-eosin and van Gieson-elastic combination stains. Additional stains such as Brown and Brenn for bacteria and Gomori's methanamine silver stain for fungi were used on selected slides.
During the period of study, 498 patients were subjected to surgery using open heart technique. In 205 of these, valve replacements were carried out using prosthetic valves in 182 and preserved homografts in 23 cases. The complication of mycotic endocarditis was encountered in this group only. The high incidence of 7.3% of mycotic endocarditis contrasts with 0.5% of pure bacterial endocarditis. In 1.5% cases fungal and bacterial infections coexisted.
In all cases except one, the definitive diagnosis of fungal infection was made only after histological examination. In one case ante-mortem diagnosis could be made on the basis of positive blood culture. In 11 cases the endocarditis developed in early post-operative period. Six of these patients died within two weeks and the rest within four weeks of operation [Table 1]. Four patients had uneventful post-operative recovery and were discharged from the hospital. They were readmitted with complications after a variable period [Table 1]. For the sake of convenience, however, the two groups are considered separately.
Both mitral and aortic valves were replaced in two cases, Mitral alone was replaced in 7 cases and aortic alone in 2 cases. Prosthetic valves were used for replacement in 8 cases and preserved aortic homografts in 3 cases. The usual signs of bacterial endocarditis, such as intermittent pyrexia, toxemia, splenomegaly and embolic phenomenon were not conspicuous in these cases. Four cases were apyrexial throughout their course. In two cases low fever was seen just for 2-3 days prior to death. In five cases there was intermittent pyrexia, but in four of these there was post-operative wound sepsis due to bacterial infection. Only in one case there was unexplained pyrexia throughout the course. No toxicity was observed in any of the cases excepting those with wound sepsis. Signs of embolization were absent in all the cases. Three patients in this group had neurological damage following cardiopulmonary bypass. They remained in semiconscious or unconscious state throughout the post-operative period and the cause of embolic lesions observed in the brain in these cases was considered to be microembolization following cardiopulmonary bypass. The only positive manifestation in 6 cases was that the cardiac function did not improve as expected without any apparent reason. Repeated blood cultures were done in all cases with intermittent pyrexia. Fungal growth was obtained only in case No. 10 just one day prior to death.
On prosthetic valves the vegetations usually started on the ring at the suture line. They gradually encroached on valve orifice producing secondary stenosis. They then gradually covered the other surface of the ring and struts and filled the cage. Occlusion of the cage was seen more frequently in aortic position than in mitral position. Only in one case the vegetations appeared to start at the aortotomy suture line invading the cage secondarily. The vegetations in fungal endocarditis usually were large and had smooth surface as compared to those in bacterial endocarditis. In fact they looked more like large thrombi and could not be distinguished from bland thrombi. In 4 cases of this group the vegetations were massive and had occluded the valve orifice to a great extent. The vegetations could grow very rapidly and near complete occlusion of the valve orifice was noted in one case within 15 days. In none of the case: there was spread by contiguity to adjacent myocardium. In case of homografts the vegetations were comparatively small, granular and were seen on the cusps. There was variable destruction of the cusps.
Embolic lesions though clinically not evident were seen either macroscopically or microscopically at the usual sites like brain, spleen and kidney in 8 cases. Interestingly micro-embolic lesions in the myocardium producing myocarditis were found to be very frequent and were seen in six cases [Table 2]. Histologically fungus was detected in emboli in three cases. After the initial experience, the vegetations in the last six cases were also cultured for fungi in addition to bacteria. In three cases (5, 6 and 10) growth was obtained and the fungus could be identified with certainty. Diagnosis on histology alone has some limitations. Essentially two types of fungi were detected in this series. The first consisted of yeast forms and pseudohyphae which were generally stained in rutine H-E stained section. This offered no problem and could be easily identified as Candida. The other type consisted of branching Septate hyphae which remained unstained in H-E stained section. For detection of this type. special fungal stains were necessary though when present in large number, it could be identified as negatively stained structures. This appearance is exhibited by Aspergiilus and Penicillium. Though there are subtle morphological differences between the two in the intensity of staining, frequency and type of branching, the diagnosis on histology alone can be fallacious. No further attempt, therefore, is made to identify the fungus and definite diagnosis is made only in cases where growth was obtained.
In contrast to cases with early endocarditis three out of four cases in this group had clinical evidence of endocarditis. Two of them had massive embolization which was the first indication of endocarditis. Prolonged low pyrexia was present in one of these but in absence of toxemia and splenomegaly the diagnosis of endocarditis was not made clinically. In the second case low fever was noted only on a few days before embolization and the patient did not seek any medical advice for that. In the third case massive embolization was the only sign of endocarditis. In the remaining case there were no indications of infective endocarditis. The patient's cardiac status gradually deteriorated and she was admitted in congestive cardiac failure.
The vegetations were massive in all the cases and had completely filled up the cage or valve orifice [Table 2]. In three cases the vegetations started on valve ring, but in the 4th case the primary site was aortotomy suture line. It secondarily filled the cage but in addition a long cylindrical thrombus 4 mm in diameter and 18 cm. in length was floating freely in the aortic lumen adherent only at the aortotomy site. Embolic lesions were present in all but massive embolization of large arteries was found in two. Histologically the fungus was Candida in two cases and in two cases septate branching hyphae were seen. Fungus was also detected in infarcts in three cases.
The high incidence of fungal endocarditis in the present series is a pointer to the increasing incidence of systemic fungal infections after prolonged and complicated surgery. Use of broad spectrum antibiotics, corticosteroids, long exposure of viscera to theatre air, tracheostomy, indwelling catheters which are part of the management of such procedure contribute towards growth of fungus, which can come from endogenous or exogenous source. The exogenous source is particularly important in tropics because of high atmospheric prevalence. As treatment of systemic mycosis was very unsatisfactory till recently, no particular attempt was made to diagnose these cases. Their increasing incidence however, warrants greater efforts in the diagnosis and management. Introduction of 5 fluorocytosine in therapeutic armamentarium has improved the prognosis of systemic mycosis. Results of treatment with 5 fluorocytosine alone or in combination with amphotericin B are encouraging. ,, Early diagnosis, therefore, would be of vital importance. The clinical profile presented by the cases in the present series amply reveals the difficulties in diagnosis. The usual indicators of endocarditis are absent in these cases. Fever if present is low and is seen generally at late stages. Toxemia is absent, splenomegaly may also be absent. Embolization is the only manifestation, but it is usually massive and fatal and does not allow enough time for remedial measures. The only positive feature is gradual deterioration of the cardiac status of the patient brought about by slow occlusion of the valve orifices. The myocardial function seems to be further jeopardized by embolic myocarditis. The diagnosis therefore would mainly depend upon laboratory aid. Positive blood culture is rare in early stages. Serial titration of antibodies against the fungi would be of greater help and rising titres would be diagnostic.  It would obviously be better to prevent than to treat the infection. In the present series there are two features which are distinct from the reported cases. The incidence of early endocarditis is very high. In these cases the fungus was introduced presumably at the time of operation or immediately after. Secondly the incidence of infection with branching septate fungus is much higher than reported series. Seelig et al , collected 87 cases of Candida endocarditis while Meyer and Fox  could collect only 13 cases of Asperpillus endocarditis during the same period. Endocarditis due to Penicillium was reported only in a single case.  Aspergillus and Penicillum are not normal habitants in human being and come from exogenous sources. Candida on the other hand may be normally present in oropharynx, vagina and intestines. Their number at these sites increases with administration of broad spectrum antibiotics. The exogenous sources need to be carefully looked for, especially the indwelling catheters and tracheostomy. The usual sites for endogenous sources also should be examined preoperatively and postoperatively to detect the colonization of fungus. The fact that fungal endocarditis was seen exclusively on replaced valves implicate heavily the valve itself as a major source. Homograft sterilization against fungal contamination is difficult as antifungal agents available here do not effectively sterilize the valves in vitro.
[Table 1], [Table 2]