‘Lipo-tourism’ gone wrong: blood transfusion-transmitted cytomegalovirus disease

Abstract

Medical tourism is the pursuit of more affordable surgeries; however, this comes at the risk of suboptimal standards and potential for life-threatening complications. In this case, we describe the diagnostic challenge of cytomegalovirus (CMV) pneumonia in a 40-year-old woman who experienced wound dehiscence and subsequent blood transfusion-transmitted CMV as complications of liposuction in the Dominican Republic. We explore the role of histopathology in the diagnosis of disseminated CMV, discuss the underlying aetiology of CMV pneumonia in this patient and weigh the risks and benefits of initiating antiviral therapy in an immunocompetent patient with CMV disease.

Background

The term ‘medical tourism’ has been used to describe the phenomenon of an abundance of Americans who travel to countries around the world seeking low-cost surgeries. Complications of these surgeries are frequently reported in case studies and series such as in the mid-2000s when Mycobacterium abscessus infections in eight American patients were linked back to surgeries in the Dominican Republic (DR).¹ According to a National Health Service study, the most common complications of these surgeries are wound dehiscence and infection, with as many as 66% requiring emergent hospital admission.²

The Centers for Disease Control’s National Center for Immunization and Respiratory Illness reports that more than half of the Americans are seropositive for cytomegalovirus (CMV) antibodies by the age of 40, though most remain asymptomatic.³ Conversely, disseminated CMV—which is associated with multivisceral disease—has traditionally been seen with conditions of immunosuppression such as HIV/AIDS, chemotherapy and the use of immunomodulating therapies in the setting of transplant. Furthermore, it is known that any state of critical illness is associated with higher rates of reactivation CMV, predisposing patients to increased duration of hospitalisation and ventilation, as well as higher all-cause mortality.⁴ Much of the existing literature regarding CMV infection is written in the context of organ transplantation. These high-risk patients often undergo pretransplant serologic testing and may in some cases receive prophylaxis to prevent systemic disease.⁵

It is important to distinguish between CMV disease, which is diagnosed by the presence of signs and symptoms, and CMV infection, which is highly prevalent but often asymptomatic.⁶ This distinction necessitates a clinical-pathologic diagnosis, one that does not rely exclusively on serology nor tissue biopsy. The gold standard for diagnosis of tissue-invasive CMV disease remains the presence of CMV inclusions on histopathology. Though there exists little consensus regarding thresholds for diagnosis of CMV disease by quantitative PCR, it is agreed on that serial testing for viral load remains the standard metric of response to antiviral therapy.^{5 7}

Symptomatic CMV can present across a spectrum of disease from mild upper respiratory or gastrointestinal illness to multivisceral, potentially life-threatening disease. The most common manifestations of disseminated CMV include colitis and meningoencephalitis, while other severe associations include haemolytic anaemia, thrombocytopaenia, pneumonia and uveitis.⁸ Treatment of CMV disease with ganciclovir and valganciclovir is often reserved for immunocompromised hosts. The data are unclear as to which immunocompetent patients with CMV disease benefit from antiviral therapy, but it is most often used in patients with severe central nervous system, ocular or pulmonary involvement.⁸ Clinicians must weigh the benefits of treatment with the risks of side effects, most notably the leucopenia, anaemia or thrombocytopaenia seen in up to 20% of patients.⁹

This case report details complications in one ‘lipo-tourist,’ who pursued panniculectomy and liposuction in the DR only to develop CMV pneumonia in the setting of blood transfusion.

Case presentation

Our case begins when an overweight but otherwise healthy 40-year-old female underwent panniculectomy with umbilical transposition and liposuction of the flanks, back and neck in the DR. She received three units of packed red blood cells (pRBC) intraoperatively and one additional unit postoperatively. Following a 2-week hospitalisation during which she received prophylactic enoxaprin, cefixime and clindamycin, the patient returned to the USA. Five days later, she presented to the emergency department noting wound dehiscence and serous drainage at the central umbilical lesion. She had not been compliant with abdominal binder or physical activity restrictions following surgery. The patient had a medical history of anxiety and post-traumatic stress disorder and remote surgeries including c-section, tubal ligation and ovarian cyst removal. The patient was taking diazepam and propranolol and had an allergy to chlorhexidine. Family history included type II diabetes mellitus in her mother. She drank alcohol rarely and had never smoked or used illicit drugs. Due to wound dehiscence complications, she subsequently underwent three excisional debridements for soft tissue infection with fat necrosis and had a vacuum-assisted closure device placed. Wound cultures were polymicrobial, growing Citrobacter, Proteus, Enterococcus faecalis, Enterococcus faecium and Staph epidermidis. She was discharged after a 16-day hospitalisation with a peripherally inserted central catheter for 3 weeks of intravenous vancomycin and ertapenem. Labs at discharge were significant for WBC 7.1, Hgb 7.8, Plt 328, and a downtrending CRP.

Two weeks later, our patient again presented to the emergency department with approximately 1 week of fever at a maximum of 103 °F as well as a worsening cough. She reported compliance with daily antibiotic regimen and had been switched from vancomycin to intravenous daptomycin 3 days prior due to neutropaenia. At the time of this presentation, vital signs showed T 37.8°C, HR 119, RR 18, BP 108/65, SpO2 96% on room air. Labs were significant for WBC 2.3, Hgb 10.3, Plt 142, AST 94 and ALT 35 (figure 1). See table 1 for complete blood count with differential at presentation. Initially, cardiopulmonary exam revealed only tachycardia with lungs clear to auscultation and a normal work of breathing. Abdominal exam showed an area of tender fluctuance at the left lateral incision site. The patient was admitted for workup.

Figure 1.

Laboratory trends in CMV pneumonia. This figure depicts the patient’s pancytopaenia, high viral load at presentation and subsequent improvement with ganciclovir induction. Treatment was initiated on hospitalisation day 6 with modest recovery of blood cell counts and an undetectable viral load at discharge. CMV, cytomegalovirus

Table 1.

Complete blood countwith differential

Patient count	Patient percentage	Patient absolute (×103/mL)	Normal range absolute (×103/mL or %)
WBC		2.3	3.4–9.6
Hgb		10.3	11.6–15.0
Platelet		142	157–371
Neutro	65.9	1.5	1.56–6.45
Lymph	23.5	0.5	0.95–3.07
Mono	7.1	0.2	0.26–0.81
Eos	0.4	0.0	0.03–0.48
Basophil	1.8	0.0	0.01–0.08

This table shows CBC with differential at the time of hospitalisation. Bolded values indicate blood counts below the normal range. The mild cytopaenia was likely the result of CMV disease.

CMV, cytomegalovirus.

Investigations

Given the patient’s history of abdominal wound infection, new onset fevers and cough, an extensive workup for infectious aetiology began including urinalysis (negative), chest X-ray (CXR) (unremarkable—figure 2A), blood cultures (negative), viral respiratory panel (negative for adenovirus, coronavirus, influenza A/B, metapneumovirus, rhinovirus/enterovirus, parainfluenza and respiratory syncytial virus) and abdominal CT revealing a new 10 × 2.5 × 6 cm abscess at the lower anterior abdominal wall. This fluid collection was aspirated, and a drainage tube was placed. The patient was treated with meropenem and daptomycin.

Figure 2.

CMV pneumonia on chest radiography. (A) CXR depicts the radiographic findings at admission versus (B) CXR at hospitalisation day 5. The repeat radiograph shows a diffuse pulmonary infiltrate consistent with CMV pneumonia. Treatment with ganciclovir was initiated the day after CXR in B due to persistent hypoxia. CMV, cytomegalovirus; CXR, chest X-ray

During the first night of hospitalisation, the patient experienced hypoxia with oxygen saturation of 88% on room air and was tachypneic with respiratory rate in the 30s. Chest auscultation revealed basilar rales, prompting repeat CXR that showed new bibasilar opacities. Subsequent thoracic CT showed diffuse bilateral ground-glass opacities suggestive of multifocal pneumonia. Further infectious workup ensued, and intravenous clindamycin was added for anaerobe coverage. Hepatitis C (HCV) screening by CIA was reactive, thought to be secondary to transfusions in the DR. However, confirmatory HCV quantitative PCR was later negative. During this workup, CMV DNA titre was found to be extremely high by quantitative PCR at 505 517 IU/mL (figure 1). CMV IgM and IgG were both positive. Historical CMV serology was not available. HIV screening was negative.

During the second night of hospitalisation, the patient experienced intermittent fevers to 39°C. On the third night, oxygen requirement increased to 3.5 L, and Hgb dropped to 6.7, requiring 1 U pRBC transfusion. Patient remained febrile and in acute hypoxic respiratory failure through day 5 of hospitalisation.

Treatment

CMV viral load had decreased to 136 942 IU/mL without antiviral therapy; however, due to persistent hypoxia and pulmonary infiltrates on repeat CXR (figure 2B), the decision was made to begin ganciclovir induction therapy at 5 mg/kg every 12 hours on day 6 of hospitalisation. All antibiotics (daptomycin, meropenem and clindamycin) were discontinued due to resolution of abdominal wall infection, increasing suspicion of a viral aetiology for pneumonia, and concerns for worsening pancytopaenia. The next day, the patient underwent bronchoscopy with bronchoalveolar lavage and transbronchial biopsy, which were ultimately negative for bacterial, viral and fungal infiltrates.

Outcome and follow-up

Despite the negative tissue cytopathology testing, she demonstrated impressive symptomatic improvement within 48 hours of ganciclovir induction with improved cough, down-trending fever curve and decreased oxygen requirement. She completed a full 14-day ganciclovir induction in the hospital with CMV viral load decreasing to 2631 IU/mL by day 12 of hospitalisation and to an undetectable level at day 17. At discharge, blood cell counts had improved to WBC 4.2, Hgb 8.4 and Plt 283 (figure 1), and liver enzymes had begun to downtrend after peaking at AST 124 and ALT 47 the day after induction. She was switched to a maintenance dose of valganciclovir at 900 mg two times per day. At 2 weeks post-discharge, the patient had made a full recovery with persistently undetectable viral load and further improved blood cell counts. Valganciclovir was discontinued at that time (figure 3).

Figure 3.

Timeline of events. The figure above depicts the timeline of events from the initial panniculectomy and liposuction in the Dominican Republic to the diagnosis and successful treatment of CMV pneumonia more than 2 months later. CMV, cytomegalovirus

Discussion

In this discussion, we seek to address three questions that arose during this patient’s hospitalisation and explain the rationale for our decision-making.

This case presented an interesting diagnostic challenge because although the patient did not have histopathologic evidence of CMV pneumonia, she experienced significant clinical improvement after treatment with ganciclovir. As discussed previously, the diagnosis of CMV disease is difficult in the absence of clear viral load cut-offs or requirement for identification by immunochemistry. Though the gold standard for diagnosis of tissue-invasive CMV continues to be biopsy and histopathologic identification, it can be made in the absence of inclusions due to a focal viral distribution that results in false negatives. According to one study on hepatic tissue samples, false negative rates are as high as 35%.¹⁰ There is also evidence to suggest that viral loads as high as 5 million geq/μg DNA may be required before the presence of inclusion bodies, well beyond the levels at which patients can be symptomatic.¹¹ Practically, decisions regarding diagnosis and treatment had to be made long before the final results of histopathology and microbial cultures. In essence, CMV pneumonia became a diagnosis of exclusion. Cough, worsening hypoxia and persistent fevers with a new multifocal infiltrates were diagnostic of pneumonia; however, the disease failed to respond to broad spectrum antibiotics nor to antifungals. Instead, our patient experienced significant symptomatic improvement within 48 hours of antiviral treatment and further improved as viral load responded appropriately to ganciclovir. PCR tests are widely used to quantify CMV DNA for this purpose of monitoring CMV disease.¹² Of note, the diagnostic bronchoscopy with bronchoalveolar lavage and tissue biopsy was performed more than 24 hours after induction therapy, at which point viral load had already decreased by two-thirds to just over 130 000 IU/mL. However, this invasive procedure was far from useless, as it decreased suspicion of an underlying fungal, bacterial or other interstitial process and added evidence in support of treatment with ganciclovir.

Was CMV disease in this patient a primary infection from blood transfusion or viral reactivation in a critically ill patient? The patient was an otherwise healthy 40-year-old at the time of the original surgery, lacking a history of cancer, HIV, organ transplant or other immunosuppressive therapy that would predispose to tissue-invasive CMV disease. The literature reports extensively on such cases of CMV disease even in immunocompetent patients and attributes most of this illness to viral reactivation in the setting of critical illness.⁴ Furthermore, a primary presentation of CMV pneumonia in an immunocompetent patient is relatively rare, which likely contributed to this patient’s delayed diagnosis and treatment of CMV pneumonia. The patient’s CMV serology prior to this hospitalisation was unknown; thus, it is difficult to distinguish between primary infection and viral reactivation based on her positive CMV IgM and IgG antibodies during hospitalisation. Though the patient could have had CMV infection prior to surgery, she most likely acquired a primary transfusion-transmitted CMV during her ‘lipo-tourism’ in the DR, where there are variable standards for blood bank maintenance. The patient was aware of multiple suspect postoperative and post-transfusion deaths from the same clinic. Whether this patient’s CMV infection was primary or reactivation of prior disease was critical not simply for the classification of our patient’s disease process, but more importantly for using evidence available to us to weigh the risks and benefits of antiviral therapy. The literature and guidelines most strongly support the use of ganciclovir in immunocompromised patients but leave room for additional research in understanding the indications for treatment in the immunocompetent.⁵

This leads us to a third question: when is it appropriate to initiate antiviral therapy in an immunocompetent patient with CMV disease? Though discussed last in our case study, this question was at the forefront of our clinical decision-making. Numerous case studies and series report the successful use of ganciclovir or valganciclovir in treatment of CMV pneumonia in immunocompetent individuals.¹³ Historically, CMV disease in immunocompetent individuals has been treated with symptomatic support rather than antiviral therapy, though there is increasing support for its wider use due to a relatively safe and effective oral option in valganciclovir.¹⁴ Our patient’s unique presentation raised additional questions that current literature does not address. To what extent should the patient’s acquired pancytopaenia have influenced the decision to induce with ganciclovir therapy, which also has the adverse effect of additional myelosuppression? Given that the viral load had already decreased substantially from 500 000 to 1 30 000 prior to induction, might the disease process have self-resolved without further complication, even without ganciclovir induction? These questions remain unanswered but should be the subject of further investigation.

Patient’s perspective.

‘After countless hours of online research, interviewing past patients, reading reviews on social media groups, and completing ongoing Q&A with his schedule assistant, I had finally settled on my surgeon… He was openly welcoming to women with higher BMI and seemingly an expert at delivering results. My decision to travel abroad has nothing to do with price. That’s a misnomer that it’s cheaper, it is not. I wanted a surgeon who was comfortable with my body type and listened to how I wanted to be portioned.

Literally, it wasn’t until the week of my scheduled surgery that he experienced the first known death due to complications. The death toll continued to mount after that. It was too late. [I was] already in the Dominican Republic, completing my pre-op testing – which included pulmonary, cardiologist screening, stress test, blood work, and dietitian – when the news surfaced. He and his staff were devastated and took time off to recoup. This delay resulted in extended surgical hours and overbooking.

My surgery was performed on a Saturday in a small facility that was reminiscent of a facial spa, not in an outpatient surgical facility. During the consultation, he determined I would need 4 bags of blood after surgery, which I considered odd, but I kept reminding myself this isn’t the U.S. After surgery, I didn’t have any complications, other than typical pain and discomfort, for 14 day in the recovery house. No one spoke English. Google translate was my life line; therefore, I kept my conversations brief and questions to minimum.

I flew home on the 4th of July. At this point, my health and life as I knew it would soon take a tailspin. I waited 8 hours in the ER, only to learn I was infected! I recounted the timeline of the weeks prior seemingly 100 times to the medical staff. The doctors seemed perplexed as to how to treat me because there were just so many unknown factors. Fast forward six surgeries later, an additional blood transfusion (totaling 5), double lung pneumonia, combined hospital stay of nearly two months, endless tests/scans, and countless IV bags of antibiotics, I was in the clear.

Due to the seriousness of my condition upon my arrival, the doctors weren’t able to be as thorough in explaining their findings nor rationale for treatment. I explained my desire to have open communication as a member of the treatment team. My voice mattered. More importantly, my comprehension and awareness of my ailments and treatment were crucial.

They listened. An on-going learning relationship of trust and mutual respect was built. Every morning, like clockwork, the team consisting of 8–10 crowded into my small hospital room with updates. Some days the news was daunting, others optimistic. Nevertheless, they assured me that they would do their best to figure it all out. And they did. I believe the transparency in communication, tenacity in research, and a collective effort to provide superb medical care are the reasons I am alive today. I’m incredibly thankful and forever indebted to the team for their hard work and diligence.’

Learning points.

• Clinicians should be aware of the prevalence of ‘lipo-tourism’ and common complications of wound dehiscence, superimposed infections and transfusion-associated illnesses.

• Cytomegalovirus (CMV) should be on the differential for multivisceral infection in any critically ill patient, whether immunocompromised or immunocompetent.

• Tissue-invasive CMV disease is a clinical-pathologic diagnosis for which treatment in an immunocompetent patient remains an important question of ongoing research.