The Perils of Precision Medicine

When The White House announced their Precision Medicine Initiative last year, they referred to precision medicine as “a new era of medicine,” signaling a shift in focus from a “one-size-fits-all-approach” to individualized care based on the specific characteristics that distinguish one patient from another. While there continues to be immense excitement about its game-changing impact in terms of early diagnoses and targeting specific treatment options, the advancements in technology, which underlie this approach, may not always yield the best medical results. In some cases, low cost approaches, based on sound clinical judgment, are still the better option. 

For example, tuberculosis (TB) is an infectious disease that continues to pose global burden with 9.6 million new cases and 1.5 million deaths reported in 2014 alone. The large toll is partly due to lack of effective treatments (particularly for drug-resistant cases) but also due to delays in diagnosis. One might think that precision medicine technology leading to improved diagnosis would be effective at minimizing the related death toll but we shouldn’t automatically assume that. It turns out that sometimes the latest technological advancements can be so sensitive that we detect organisms that are not causing disease.

I recently took care of a 50 year old man who was admitted to our hospital with fever and cough. Not a healthy man, he had liver cirrhosis, chronic obstructive pulmonary disease, and substance abuse. He also had a history of latent TB infection based on a positive TB skin test, but he never had active TB manifesting with clinical symptoms.   He never received prophylactic TB medications to prevent the development of active disease. Because of this history, his doctors sent a sputum sample for TB cultures, which usually take 6 weeks to grow the organism. The patient went home since he felt better with treatment for bronchitis.

In the meantime, the laboratory reported that the patient’s sputum cultures were growing a “TB-like” organism.  The laboratory tested this culture using Gene Xpert, a precision medicine tool developed in 2010, which yielded a positive identification of drug resistant TB (along with non-TB mycobacteria). The patient was readmitted for TB treatment. Although the patient informed his doctors that he felt better, the question of active TB persisted.  Prior to treatment, his sputum samples were re-tested and all were negative using GeneXpert and traditional culture tests. The laboratory decided to further use Next Generation sequencing, another tool of precision medicine to test the original sample which had the positive identification. This test showed that a few TB bacteria were present and genomic testing showed the presence of drug-resistance.

The doctors decided to treat the patient with multiple medications for multi-drug resistant TB.   Over the next six months, the patient received an array of antibiotics that produced toxic side effects involving his kidneys, liver and lungs. The doctors conferred with the state health department and it was decided to continue treatment largely on the basis of the testing and the concern for potential infectiousness to the public.  When I became involved in this patient’s care, he was doing poorly and I wondered if we were doing the right thing. 

After all, the patient did not show classic signs of active TB (fever, bloody cough, weight loss).  Yet, our precision medicine tools revealed that resistant TB, albeit few in number, were present in his sputum. But did this constitute disease?  In infectious disease, there has always been the concept of “colonization,” where bacteria live on the human body but do not cause disease. But does this concept apply to TB?  The natural history of this infection describes a spectrum of TB ranging from harboring the bacteria in latent form to actual full-blown symptomatic disease when bacteria reach a critical number.  As our tests get more sophisticated, we may pick up earlier and earlier on conditions that do not actually represent disease as defined by the presence of symptoms.  But is there a benefit (to the patient and the public) in treating earlier in the absence of symptoms? Perhaps, but the more we treat, the more potential side effects can ensue.

In the end, our patient suffered multiple toxicities and his quarantine made him “feel like a prisoner.” Eventually, after consultation with local, state, and federal experts, including an institutional ethics consult, we decided to discontinue TB treatment in favor of monitoring him closely as an outpatient.  He died 6 months later.

I’ve always wondered if our prolonged attempt at TB treatment hastened his demise.  He never did show any signs of active TB prior to his death.

The Hippocratic oath is credited with the concept of primum non nocere or  “first do no harm.”   Further stated, physicians pledge that “I will according to my ability and judgment, prescribe a regimen for the health of the sick; but I will utterly reject harm and mischief.” To be sure, new precision medicine technologies can reset the clock on when we can first detect causes of disease and offer guidance on what types of treatment we can use. But how do we figure out how to draw the correct line between health and mischief?  As clinicians, we must learn to balance the promises of technology with a practical wisdom so that we do not put our patients in harm’s way.

The author is director at the AIDS program Yale School of Medicine.

Categories: Uncategorized