A comparison of the surgical outcomes of ventriculoperitoneal versus lumbar peritoneal shunts in the management of intracranial hypertension secondary to cryptococcal meningitis in HIV infected adult patients

Objective: Cryptococcal meningitis (CM), an AIDS-defining illness, significantly impacts morbidity and mortality. This study aims to compare complications arising from ventriculoperitoneal shunt (VPS) and lumbar peritoneal shunt (LPS) procedures used to manage refractory intracranial hypertension (IH) secondary to CM in HIV-infected patients. Methods: Retrospective data were collected from January 2003 to January 2015 for HIV-infected adults diagnosed with refractory IH secondary to CM and subsequently shunted. Demographics, clinical characteristics, antiretroviral therapy, laboratory findings (including CD4 count and CSF results), CT brain scan results, shunt-related complications, and mortality were compared between VPS and LPS groups. Results: This study included 83 patients, with 60 (72%) undergoing VPS and 23 (28%) receiving LPS. Mean ages were comparable between VPS (32.5) and LPS (32.2) groups (p = 0.89). Median CD4 + counts were 76 cells/ µ l (IQR = 30 – 129) in VPS versus 54 cells/ µ l (IQR = 31 – 83) in LPS (p = 0.45). VPS group showed a higher mean haemoglobin of 11.5 g/dl compared to 9.9 g/dl in the LPS group (p = 0.001). CT brain scans showed hydrocephalus in 55 VPS and 13 LPS patients respectively. Shunt complications were observed in 17 (28%) VPS patients versus 10 (43.5%) LPS patients (p = 0.5). Patients developing shunt sepsis in the VPS group exhibited a median CD4 + count of 117 cells/ µ l (IQR = 76 – 129) versus 48 cells/ µ l (IQR = 31 – 66) in the LPS group (p = 0.03). Early shunt malfunction occurred more frequently in the LPS group compared to VPS group (p = 0.044). The mean hospital stay was 6.2 days for VPS versus 5.4 days for LPS patients (p = 0.9). In-hospital mortality was 6%, occurring in three VPS and two LPS patients respectively. Conclusion: Shunting procedures remain important surgical interventions for refractory IH secondary to HIV-related CM. However, cautious consideration is warranted for patients with CD4 counts below 200 cells/ µ L due to increased shunt complications. This study suggests a trend toward higher complication rates in patients undergoing LPS insertion.


Introduction
Central nervous system (CNS) infections are associated with increased mortality and morbidity in individuals living with HIV, particularly in low-and middle-income countries (LMIC) [1].In South Africa (SA), 7700,000 people were reported to be living with HIV, with a prevalence rate of 20.4% among the population aged 15-49 years [1].The Province of KwaZulu-Natal (KZN) has been the most severely affected, with the highest prevalence rate at 26.8% [2].
Cryptococcal meningitis (CM) is one of the most common CNS opportunistic infections diagnosed in people living with HIV.CM is regarded as an AIDS-defining illness, diagnosed in individuals with latestage HIV infection.Sub-Saharan Africa and Southeast Asia report annual CM prevalence rates ranging between 12% and 50% and 6-18%, respectively [3][4][5][6][7][8].
Currently, there is no consensus on the best form of CSF shunting procedure in this patient cohort.Shunt-related complications must be taken into consideration when selecting suitable candidates for shunt surgery, especially in patients with a CD4 count below 200 cells/uL.The purpose of this study was to compare the complications of these two types of CSF procedures in this group of patients.

Methods
A retrospective chart review was conducted including all adult HIVinfected patients (18 years and above) diagnosed with refractory IH secondary to CM who underwent either a VPS or LPS procedure in the Department of Neurosurgery (DoN) at Inkosi Albert Luthuli Central Hospital, Durban, South Africa.The study period was between January 2003 and January 2015.
Data collected and compared among HIV-infected patients who underwent these shunting procedures included demographics, clinical presentations, Glasgow Coma Scale (GCS) at admission, lumbar puncture opening pressure (LPOP), HIV-associated co-morbidities, CD4+ count levels, antiretroviral therapy status, presence of hydrocephalus/ dilated ventricles on CT or MRI brain scans, CSF, hematological and biochemical findings.A comparison was drawn between VPS and LPS procedure outcomes in terms of shunt survival, complications, hospital stay length, and in-hospital mortality.
CM was managed following standard guidelines [18].IH was defined as lumbar puncture opening pressure (LPOP) greater than 25 cm of water (H2O).CT brain scans were performed to assess ventricle size, focusing on the presence or absence of radiological features of hydrocephalus features.Refractory IH was diagnosed when persistent increased intracranial pressure (ICP) endured for more than 12 days, despite repeated daily therapeutic lumbar punctures (LP), associated with decreased level consciousness, worsening neurological deficits, visual deterioration, and difficulty in tolerating LPs.
Exclusion criteria encompassed HIV-negative patients diagnosed with CM related IH and those under 18 years of age.Ethical approval was granted by the Biomedical Research Ethics Committee of the University of KwaZulu-Natal (Ref No: BE060/18).

Statistical analysis
Descriptive statistics summarized patients' demographic and clinical characteristics.Frequency distribution of numeric data was assessed for normality, presented using means and standard deviations for normally distributed variables and medians with interquartile ranges (IQR) for non-normally distributed data.Chi-square tests identified categorical factors associated with shunt complications, including sex, clinical characteristics, antiretroviral therapy, and CT brain scan findings such as hydrocephalus and infarcts.
T-tests or Mann-Whitney tests were used to compare numeric risk factors such as age, admission GCS, LPOP, CD4 count, CSF, hematology, and biochemical results between VPS and LPS groups.Kaplan-Meier survival analysis reported VPS and LPS survival and length of hospital stay (LOHS) between the two groups.Stata v13 was used for data analysis, setting the significance level at 0.05.

Results
During the study period, a total of 1382 HIV-infected patients received treatment for various pathologies in the Department of Neurosurgery (DoN).Among them, 83 (6%) patients underwent shunt procedures following the diagnosis of refractory IH secondary to CM. VPS procedures were performed on 60 patients (72%), while LPS procedures were performed on 23 (28%).Table 1 outlines the demographic and clinical characteristics comparison between the VPS and LPS groups.
Shunt complications arose in 27 patients (32.5%), with 17 cases (28%) in the VPS group and 10 cases (43%) in the LPS group.Among these, shunt sepsis occurred in eight cases (47%) in the VPS group and eight cases (80%) in the LPS group, with the remaining complications attributed to shunt malfunction (VPS: 53%; LPS: 20%).Key factors associated with shunt complications in both groups are presented in Table 3. LPS infection occurred at a significantly lower CD4 count when  compared to VPS infection (p=0.003,Table 3).
Comparative survival analysis between VPS and LPS is depicted in the Kaplan-Meier graph (Fig. 1), revealing early shunt malfunction in the LPS group compared to the VPS group.
Staphylococcus epidermidis (12 cases; 75%) and Staphylococcus aureus (4 cases; 25%) were identified as the primary organisms responsible for shunt sepsis.Anaemia was more prevalent in the VPS group compared to the LPS group (p=0.02,Table 3).
The mean length of stay was 6.2 (8.5) days in the VPS group compared to 5.4 (4.2) days in the LPS group (p=0.90,Fig. 2).During the admission period, five patients (6%) died, three (5%) from the VPS group and two (9%) from the LPS group.However, the sample size for mortality was small for statistical analysis.

Discussion
To our knowledge, no study has compared the complications of VPS versus LPS following management of refractory IH secondary to CM in HIV-infected adult patients.In our study, males were more affected than females, similar to reports by other authors [13,17].HIV infection in South Africa has been reported to affect more females than males [18][19][20].The mean age group reported in our study further reinforces that the HIV pandemic has affected mainly young people.
In our study, most patients presented with late-stage immunosuppression, with a median CD4+ cell count of 76 cells/ul in the VPSI group and 54 cells/ul in the LPSI group.The World Health Organisation (WHO) defined late-stage immunosuppression as a CD4+ cell count of less than 200 cells/µl [6].Baddley et al. reported similar findings in their study [17].CM results in chronic granulomatous meningitis, characterized by diffuse lymphocytic meningitis along with foci of inflammation in the basilar meninges [21,22].
In the setting of AIDS, with absent host cellular immune response, CM follows a trajectory closer to a massive fungal infestation with overwhelming numbers of cryptococci in the CSF which obstruct the arachnoid villi directly [9,23].This may explain why meningism was found in two-thirds of our patients and why headache was the most common clinical feature in this population.
Sixteen (18.3%) of our patients presented with blindness, which is a debilitating complication and often is not reversible.Lui et al. found that blindness caused by optic nerve damage because of intracranial hypertension may be reversible following diversion of CSF [16].Forty-nine (59%) patients were on ART, and this must be viewed in the context that in 2002, ARTs were not freely available to patients in SA.Access to ART has improved over the years in SA with quoted figures of 59 000 patients on ART in 2005 and 144 000 patients on ART in 2015 [20,24].
In our study, dilated ventricles were found in 68 (82%) patients followed by normal-sized ventricles in 15 (18%).This finding is important because CM can result in intracranial hypertension in the absence of dilated ventricles.This is due to the coating of the ependymal tissue by cryptococcal capsular polysaccharides leading to the failure of the ventricles to dilate.Our findings, however, differed from Tan et al. who found that despite elevated LPOP, most patients did not demonstrate dilated ventricles on CT scans [25,26].
Dilated ventricles on CT brain scans were an important factor in our unit when determining the type of CSF shunting procedure to be performed.When the ventricles were dilated, the treating neurosurgeon preferred VPSI as the ventricles were easy to cannulate when compared to normal-sized or small ventricles, where LPSI was preferred.However, navigation has made it easier and safer for neurosurgeons to insert VPS even in the setting of small ventricles.A study of 49 patients by Yim et al. found that the use of neuro-navigation in the placement of ventricular catheters in patients with small ventricles achieved a 90% (n= 44/49) correct placement position [27].
Most studies reporting on surgical management of IH associated with CM used VPSI [13][14][15][16][17], showing a trend of preference for VPSI over LPSI   by most authors.Our study showed a higher rate of shunt complications in the LPSI group versus the VPSI group.This finding is in keeping with reports by Menger et al. who found that LPSI has a significantly higher failure and revision rate of 7% when compared to a VPSI rate of 3.9%.They also reported a longer length of hospital stay in the LPSI group when compared to the VPSI group [28].
Studies comparing VPSI versus LPSI in the treatment of IH, tend to favour the former over the latter due to higher complication and revision rates in the LPSI groups [28][29][30].Abubaker et al. found a higher revision rate of 60% in the LPSI group compared to 30% in the VPSI group [29].Tarnaris et al. found a 24% complication rate in the LPSI group compared to 11% in the VPSI group [30].
Sharma [31] and Choksey et al. [32] found that shunt infection rates in HIV-infected individuals were unacceptably high.The overall rate of VPS infection in non-HIV infected individuals is reported at 1-5%, however in HIV-positive individuals, this ranges between 5% and 15% and some authors have described higher rates at 7-12% [32].
A study by Calvo et al. showed that patients with reduced immunity are at a higher risk of post-operative infection than nonimmunocompromised patients [33].A similar finding was made by Woodworth et al. [34].Reasons for reluctance to shunt these patients include higher rates of colonization of the shunt, the theoretical risk of seeding of the fungal spores into the peritoneum and shunt obstruction due to high titres of fungal polysaccharide [35].
The most common isolated organisms among our study population were Staphylococcus epidermidis and Staphylococcus aureus.Staphylococcus species is by far the most common organism responsible for shunt infections accounting for up to 90% of all isolated organisms [32,[36][37][38].A potential option in mitigating this complication is with the use of antibiotic-impregnated shunts (AIS) [38].A mortality rate of 6% was found in our study which was less compared to the mortality rate reported by Cherian et al. [13].

Limitations
As a retrospective study, management decisions are assessed after the fact, and bias in the treating surgeon cannot be accounted for.Patients referred to the DoN come from various regional and district hospitals all over the province and upon discharge, continuous follow-up can be a challenge.
This study focused on patients in the public sector and did not include the private and therefore might not give a true reflection of the overall prevalence and outcomes of patients managed in KZN.

Conclusions
Shunting procedures remain important surgical interventions for refractory IH secondary to HIV-related CM.However, cautious consideration is warranted for patients with CD4 counts below 200 cells/µL due to increased shunt complications.In our study, there was a trend toward higher complication rates in patients with low CD4+ cell counts and those undergoing LPS insertion.

Fig. 1 .
Fig. 1.Kaplan-Meier graph demonstrating shunt survival in the VPS group (Blue line) versus the LPS group (Red line).The log-rank test was used to analyse data and showed early shunt malfunction in the LPS group compared to the VPS group (p=0.044).

Fig. 2 .
Fig. 2. Kaplan-Meier graph demonstrating length of hospital stay in the VPS group (Blue line) versus LPS group (Red line).The log-rank test was used to analyse data and showed the mean length of hospital stay was 6.2 ±8.5 days for VPS group compared to 5.4 ±4.2 days in the LPS group (p = 0.9).

Table 1
Comparison of demographic profile and clinical characteristics of HIV infected patients shunted for refractory IH secondary to CM.

Table 2
Comparison of laboratory results of HIV infected adult patients shunted for refractory IH secondary to CM.

Table 3
Analysis of factors associated with VPS and LPS complications.