- Cardiology: Electrocardiogram (EKG)
- Both American and European cardiological societies (Circulation 2007, 116:2634-2653) are quite clear on this: myocardial infarction is defined by an elevation in cardiac markers, particularly troponin, with suggestive symptoms (i.e. chest pain) or electrocardiographic changes (actually that's just one definition there are others). Specific EKG changes such as new ST-elevations or new LBBB need acute cardiac catheterization as does non-ST elevation MI with continued pain. The argument that because the EKG is normal or "nonischemic" that this presentation is not related to cardiac ischemia cannot be validated until subsequent cardiac markers are negative, four to eight hours later. Although if you do suspect cardiac disease, check them sooner as you may be able to detect a positive trend. Remember too that an EKG is a snippet of time, if the patient is asymptomatic at the time, the associated electrocardiographic event may never be recorded on the EKG, particularly in cases of paroxysmal arrhythmia. It just goes to show that an EKG with diagnostic findings need to be corroborated with further diagnostic testing and that a pristinely, normal EKG has about as much association with mortality as tea leaves do. It also shows that telemetry, if it can be correlated with episodes of symptoms,` may increase or decrease the likelihood ratio of a cardiogenic etiology.
- Critical Care: Arterial blood gas (ABG)
- The ICU is essentially a place for one or both of two things: ventilators and vasopressors. Thus if the patient is in some form of shock to the ICU they go. You don't need an ABG to diagnose shock, although a PaCO2 < 32 mmHg is one of the systemic inflammatory response criteria, and would be useful but not necessary to diagnose sepsis. Elevated lactate > 2 mmol/L and not the pH is criterion for severe sepsis. The decision to intubate can and should be made independent of the blood gas (Manual of Emergency Airway Management), how the patient appears clinically in terms of work of breathing, airway protection, and level of consciousness is a better indicator of requiring invasive positive pressure ventilation. The ABG allows you to assess response to mechanical ventilation and to make adjustments that the patient's body is, hopefully transiently, physiologically unable to do. Thus the arterial blood gas is a powerful tool for assessing how a patient is currently doing and how they have responded to interventions, it does not clinch the diagnosis of shock nor is it the tool to decided whether or not a patient should be intubated.
- Endocrinology: Cortisol
- No laboratory result is more nebulous than the cortisol level. Although a random cortisol level < 3 ug/dL is highly suspicious for adrenal insufficiency and > 10 ug/dL is unlikely to be adrenally insufficient, further testing using ACTH stimulation is often required to delineate if they actually have adrenal insufficiency. To the endocrinologists chagrin corticosteroids have usually already been given, dexamethasone being the one that shouldn't confuse further diagnostic testing. The amount of ACTH to administer and the time to wait superimposed on the reality that the timing of drug administration to lab draw is seldom accurate makes the subsequent determination of the result a question fraught with perils, particularly when the results make adrenal insufficiency less likely but the administration of stress dose corticosteroids resulted in a marked clinical improvement. This may be why the administration of corticosteroids without checking serum cortisol level is part of the Surviving Sepsis Campaign.
- Gastroenterology: Rectal exam and fecal occult blood testing
- The rectal exam is the most avoided exam there is, the reason for which is presumably related to provider discomfort and the time it takes. It is only clinically important when it is not done. The rectal exam is touted in the work-up of gastrointestinal bleeding, but is only useful if frankly positive, i.e. active bleeding is observed, indicating that bleeding is most likely rapid and/or in the distal colon. A negative fecal occult blood test does not eliminate gastrointestinal bleeding. A positive test indicates that there is blood in the gastrointestinal tract the source is unclear and swallowed blood from another source is not excluded. In fact Intern Med J. 2010 Feb;40(2):107-11 claims that "there is no place for FOBT in an acute hospital setting."
- Hematology: Peripheral smear
- The peripheral smear are the tea leaves of the hematological world. For the work-up of too much or too few platelets, red, or white blood cells it is the go to test of choice for board examinations, usually with an attached image for our interpretation. While diagnostic in some conditions and helpful in directing clinical decision making in others it is no longer a convenient test. A long time ago in a hospital far, far away residents and fellows performed peripheral smears on the floor. They drew their own blood, prepared their own slides, and viewed them under the microscope. They had clinical information within minutes. In today's hospital peripheral smears are prepared in the lab, reviewed by the pathologists every second Thursday of the month, and finally obtained by the hematologist for a second look. That makes this test more mysterious to the generalist because we don't do it and less diagnostically useful due to the time lag.
- Infectious Disease: Gram stain and culture
- The potions and poisons of infection are a world with constantly evolving diagnostic results, starting with the Gram stain, followed by the quantitating and qualifying the bacterial type, and finally determining the susceptibility to antibiotics. Testing after antibiotic administration nullifies the accuracy of results, and some infections leak so few bacteria, i.e. endocarditis, that not two sets but three sets of blood cultures are drawn. Unfortunately empiric therapy must be initiated prior to the test results, and even if those results are negative, patient's often improve while antibiotics are on board whether due to them or not is unclear. Antibiotics have a recommended duration, three days for uncomplicated urinary tract infection, and weeks for osteomyelitis, endocarditis and line infection. Once committed to a diagnosis we are committed to a course and the attendant complications of multi drug resistance and C. difficile diarrhea. Oft touted clinical markers of infection such as fever or leukocytosis are not exclusive to infection. As we grow older our febrile response weakens, meaning that an elderly person without a fever can still be septic and bacteremic but treating the marked leukocytosis of CLL with antibiotics is well amusing to our specialist colleagues.
- Nephrology: Urine analysis
- The urine analysis, the tea leaves or at least the result of them, is the bedeviling test of renal disease. Unfortunately the urine analysis is nonspecific for severity of disease. Furthermore the results are affected not only by what portion of the stream the urine is obtained but the time between acquisition and analysis. The versatility of the urine analysis decreases the worse the renal function becomes. Woe to he or she who obtains the UA in ESRD while in the anuric patient the urine analysis impatiently awaited by the nephrologist isn't going to change anything, there's no pee for them to see.
- Neurology: Magnetic resonance imaging of the head
- Unless of course the patient has the wrong type and collection of metal in the the wrong location in which case the test of choice is the CTA. Prior to tomographic imaging, neurology used carefully elicited signs to diagnose the diseases neurological. Thus neurologists nearing retirement these days amuse themselves by diagnosing the disease and pinpointing the lesion (Lateralize! Levelize! Localize!) before the imaging can get done. They then sit back and chuckle at our wonder at their acumen. With the advent of three dimensional brain and spine imaging, the need for superior exam skills decreased and reliance on the imager for diagnosis has increased.
- Oncology: Biopsy
- Despite the cases where endoscopists and radiologists have put their reputation on the line by saying, prior to any pathological data and with minimal clinical data, "looks like cancer", heavens be merciful to anyone who pages the oncologist with the right story and subjective evidence but no pathology. Unless the pathology is back, after being confirmed at a tertiary center, it is not cancer to the cancer doc! In their defense, no cancer will kill before the pathology is back, if it does the disease was so advanced chemotherapy would not have made a difference anyway. Also treatment regimens are titrated to the diagnosis, so pathology is important for directing therapy and prognosis.
- Pulmonology: Pulmonary function test (PFT)
- In the diagnosis of obstructive versus restrictive lung disease and for the establishment of chronic obstructive pulmonary disease (Global Initiative for Chronic Obstructive Lung Disease) the pulmonary function test is key. However they are performed on patients at their baseline to obtain measures of severity of chronic disease and aren't usually available in-patient. Also PFTs never get better, given the limited therapeutic options we have for pulmonary disease, we don't manage the disease so much as try to decelerate its decline. So with the knowledge they have a chronic problem, we have little evidence to know who will need merely some nebulizer treatments, who will need intubation, and who will die. But still the pulmonologists will ask for or refer back to the PFTs obtained at a time of relative pulmonary stability.
- Rheumatology: Inflammatory markers and nearly anything prefaced with anti-
- When the signs and symptoms make no sense, the cause is either rheumatologic or psychogenic. Autoimmune and connective tissue disease are diagnosed with inflammatory markers and multisyllabic serum tests starting with anti- (i.e. anti-double stranded DNA antibody or anti-phospholipid antibody). After a battery of these tests, which are oft repeated because of their poor sensitivity and specificity the rheumatologist will arrive at some equally multisyllabic diagnosis (i.e. systemic lupus erythematosus or granulomatosis with polyangiitis (Wegener's)). And then they prescribe corticosteroids.
Friday, November 4, 2011
Friday, October 21, 2011
People should be admitted to the hospital because they are sick enough to have an increased chance of dying relative to those going home. My overall patient-by-patient "mission statement" is: Find the diagnoses and offer treatment options that decrease acute mortality and chronic morbidity through efficient and high quality care. Offer palliative care options when such treatment is medically futile or does not provide significant gains in the quantity to quality of life balance. The patient needs to be looked at as a gestalt of both data and intuition regarding the level of acute injury inflicted on the substrate of a certain level of chronically damaged protoplasm. I have to explain to my patient and their circle that based on what I think is going on, will the treatment options available improve their lives and do those benefits outweigh the risks of that treatment. If not, pursuing aggressive treatment of the untreatable is a study in futility, and palliative care options should be introduced. Understanding the worst case scenarios, is in my opinion, more valuable to both provider and patient than painting a rosy picture of the future. It is important to foresee negative outcomes to know what to test for and what you may need to treat, but without either you or the patient losing sight or hope of positive outcomes. Positive outcomes do not necessarily mean a cure, it may mean pain control, dignity, and even a peaceful death.
Sweat the Details
A good internist is a scared internist. Worried about what has happened, concerned about why the current state of their patient is different from before, wondering how complications can occur in the future. We take a microscope to abnormalities, trying to tease out if this is incidentally benign or malignant, wondering if this is evidence of disease or simply a normal perturbation. Hospitalists must be the diagnostic laser, making lightspeed decisions about what to pursue and what to defer as well as carefully convey to outpatient follow-up. We must carefully dissect subjective complaints and objective data using the scalpel of evidence-based criteria and a wealth of subspecialty guidelines. We must thin the electronic medical record of redundant or inaccurate diagnoses while updating it with new history. The patient's medication list is not simply the medications they are prescribed but those they are actually taking and when they are taking them. A hospitalized patient demands an agent, their hospitalist, who manages nutrition, hydration, electrolytes, medications, and ancillary therapist evaluations while requesting appropriate specialist consultation and preventing adverse events such as deep vein thrombosis, falls, and delirium. We must take a collection of prior documentation to weave a history, amalgamate in-patient events, tests, and consults as daily progress notes, and then meticulously sift all of this into a timely, complete discharge summary.
Achieving Hemodynamic Homeostasis
I consider vitals to be an acronym for Verify If They Are Living Still. In general most hemodynamically regular people are euthermic with a heart rate between 60 and 90 beats per minute, breathing 8-16 times per minute, have a systolic blood pressure greater than 100 mmHg and less than 160 mmHg, and saturate greater 92% on room air. If they don't they are either (1) acutely sick with a physiological response, (2) chronically sick and at their (albeit abnormal for the majority of the human population) baseline, or (3) weird. Acutely ill humans should have a positive response to appropriate therapy, that is there vital signs approach normal, and negative one to inappropriate therapy. That being said, people die even despite the most optimal of treatment decisions, and people live despite our best unintentional attempts at killing them. You have to determine whether the population mean abnormal vital sign(s) are an indicator of acute pathology or simply evidence of chronic but stable disease.
Returning the Patient to Basal Levels Lab Result Rubor
People are often admitted because of the density of red (abnormal) test results they have. Although abnormals should be recognized they should also be placed within the context of their chronicity and severity. Some patients have bloody diagnostic tests at baseline. An elevated white blood cell count can be due to chronic lymphocytic leukemia, dehydration, stress, or infection. No matter what we as physicians do, it is unlikely that no matter what we do acutely we cannot change the leukocytosis of CLL. Antibiotics do not treat dehydration or stress, but not treating a pneumonia with its associated leukocytosis is tantamount to medical crime. Similarly a volume depleted patient may have a CBC in the black, volume contraction hiding their basal anemia, just because they are less in the red, doesn't mean they are any less sick. An elevated creatine in a patient on dialysis has an impressive crimson hue but generates little therapeutic inertia compared to the same value in someone without previous kidney disease. No matter what intervention is undertaken, the patient with end-stage renal disease will never be in the black when it comes to creatinine, if we waited for this value to normalize before discharging the patient, they would serve a life sentence. As hospitalists we must find with the intent to treat the scarlet letter of abnormal but accept the chronic unchangeable crimson that will show up on some diagnostic tests. We can only attempt to restore our patient to the baseline they had before and hope to prevent the next decompensation as best we can.
Saturday, October 8, 2011
I've been a professional student for (too) many years, but the education that I took for granted ended the day I started as attending. The dreaded morning report hosted by harried if dedicated clinicians, the complex grand rounds elucidated by leaders in medicine, noontime and core conferences in a multitude of subjects that were a constant time pressure in residency were rudely stolen the day I graduated. Within months of being a junior attending I was academically starved. Thus I started looking for other opportunities, seeking out the rounds I used to loath, traveling to conferences in far cities, and looking up events that just sounded interesting.
This week I attended the Illinois Transdisciplinary Obesity Prevention Program's (I-TOPP) Inaugural Biennial Symposium, which covered factors from the microscopic to the transnational of how obesity can be prevented in children. I'm an adult nocturnist, the youngest patient's I see are 19, and the only preventative medicine I do is of a more acute nature, as in preventing my patients from stopping oxygenating or perfusing. What's the point of going to something that I would be unlikely to have any "need" of. Knowledge isn't like that, it is insidious, flashes of brilliance (or snippets of insanity, depending on your point of view) occur while learning the most unlikely and seemingly unrelated things.
The symposium was opened by an epidemiologist at the Centers for Disease Control, Dr. Cynthia Ogden, who went through the data of childhood obesity and problems with the precision and accuracy of that data. Obesity in children is a huge problem, but may fortunately have plateaued. However those children are an added source of patients to an already growing cohort of obese older adults.
Obesity is often simplified as energy in (EIN) greater than energy out (EOUT), which to most clinicians, including myself, means eat fewer calories and expend more. However as I learned, this is a gross oversimplification, not of the equation, necessarily, but of the complex and multitudinous variables that contribute to both energy in and out, including (but certainly not limited to) calorie source, quantity, timing of consumption, genes, gut microbiota, being breastfed, psychosocial factors, location, economics, and many more. These variables are further defined by confusingly dependent and independent equations that have yet to be defined.
Genes! Impossible! Not according to Dr. Molly Bray, who examines genes in obesity. She told us about her and others research into genes that not only affect metabolism but behavior driving the obese to not only have decreased satiety but to shy away from exercise. Phenotypically these genes are environmentally dependent. In times of food scarcity they are genotypically pervasive but phenotypically absent, however when an "obesiogenic" environment presents itself, so too does the obesity phenotype.
One of the genes that has been discussed extensively is fat mass and obesity-associated protein (FTO). FTO encodes for proteins which can demethylate DNA, or promote gene products. This is an amazing concept, since dietary choices, if they are indeed choices and not mandated by genetic and environmental factors, produce concentration gradients of carbohydrates, lipids, and proteins which can interact directly with DNA, and therefore for may affect FTO and therefore drive demethylation. My conjecture is that this then in turn could produced a cascade effect on other genes involved in metabolism, cellular protein production, and behavior that would drive an organism toward obesity with different velocities. You are not only what you eat but you are regulated by it as well. It is as if the lumber and mortar got to to be the building supervisor as well.
The next invited speaker, Dr. Madeleine Sigman-Grant, discussed the development of the All 4 Kids Program, a community-based intervention in pre-schoolers to prevent obesity designed using a Logic Evaluation Model. I don't deal with preschoolers in a professional capacity nor am I planning any community interventions anytime soon. But it was a rewarding and interesting experience to see how someone worked through a Logic Evaluation Model to develop a preventative health intervention.
Dr. Stephen Matthews spoke on spatial polygamy, which is less risque than it sounds. The concept of spatial polygamy refers to individuals having multiple spatial interactions depending on the map technique used. Thus we are defined by zip codes, voting districts, census demographics, street addresses, telephone numbers, and e-mail addresses that redefine associations between individuals and their spatial or temporal dependence. Thus it is important to understand the source of our mapping data and the limits thereof.
This creates an interesting parallel in hospitalized patients. Their physical location in the hospital has little to do with their "therapeutic distance" from their physician. Therapeutic distance would be defined as the time between a patient having a need and it being taken care of. Two patients in the same room may have vastly different care demographics and therefore therapeutic distance, even if they are physically adjacent. The patient with the in-house hospitalist versus the traditional internist on-call from home may have significantly different response times to pages. The comanaged patient may or may not be more therapeutically isolated depending on how clear the nursing staff is on the division of duties between the physicians caring for the patient, i.e. if they don't know who to call, they will take longer to get the right physician to solve the problem.
Despite not being "my area" I certainly took a step outside my box, which would have expanded to include this new information, thus creating a new and larger box. Sometimes the box doesn't change but leaping outside the box in one area just places us within familiar and useful territory in another. Each day of our lives should be a leap outside of our old box, but within a new and larger box. The only way to achieve that is through the core competency of lifelong learning, that is no longer quite such a continuous feed after entering the post postgraduate world.
Sunday, October 2, 2011
A tongue in cheek application of the Laws of Thermodynamics as applied to medicine:
0th: If two diagnostic findings support a third, then they must support each other. No single diagnostic finding has ever made a diagnosis, it is a careful process of combining and comparing data that makes it possible to reach a diagnostic conclusion. Look at what your "facts" are, the diagnosis with the most "facts" supporting it is most likely the right one. Of course, facts in medicine are not as concrete as we would like to make them seem as demonstrated by false positive and negative results.
1st: You cannot win — A physician's diagnostic performance can be of one form (i.e. specific or sensitive) or another, but cannot be both. If you rule out a lot of different etiologies without ever ruling one in you are highly sensitive but not specific, and your patient still doesn't know what is wrong with them. If you always get positive results to your diagnostic inquiries you are highly specific but not very sensitive, and you are probably missing some disease in other patients because of missed testing.
2nd: You cannot break even — A negative diagnostic work-up does not decrease morbidity or mortality, while a positive one always increases both. Just because you didn't find a pathophysiological reason for a presentation doesn't mean that you didn't miss something or that your patient is any safer from future insults. Confirming a diagnosis simply means that you have named the disease and they now have the same increased risks as others with that disease as well as those from related or unrelated future insults.
3rd: You cannot leave the game — Absolutely false results must be worked-up, as there are no absolutes in medicine. Even the most spectacularly sound medical reasoning for explaining a false positive or random incidentaloma means nothing without further testing to corroborate that reasoning. Missed diagnoses kill, maim, and get doctors sued.
Friday, September 16, 2011
I Injury (caused by an exogenous agent, this would include hemorrhage)
I Inflammation (caused by the human organisms response to injury)
Y Immune (caused by under, over, or inappropriate response of the human self-defense mechanism)
I° Infection (caused by an exogenous microbial vector)
I Ischemia (caused by lack of blood flow)
i Inadequacy (caused by mechanical failure)
Aye Incompetence (caused by physiological failure, heart failure would presumably fall in both)
I Intoxication (caused by poison)
I Iatrogenic (caused by doctors)
¿ Idiopathic (caused by an unclear reason that is not included in the above, the cop-out)
Thursday, July 14, 2011
Let us start our tale with little Ms. P who has a relationship with the broad shouldered Mr. QRS. In a normal stable relationship (NSR), Ms. P is always accompanied by Mr. QRS. They plan to go to weekly dance classes with some other partnered friends, but it's difficult with the sometimes early and other times late Sinus Arrhythmia. The Sinus Exit Block couple are always punctual when they make it, but sometimes skip a dance class or two with no warning, only to be there punctually next time. The hardest folks to do any planning with is the Sinus Arrest couple, they not only miss classes they never get the time right on the next one, arriving early or late.
Relationships can be challenging, like when Ms. P's ex-boyfriend Pac (Premature Atrial Contraction) drops in. He wants to hang out with P and reminisce about old times, messing up the lovebirds orderly schedule. The same thing happens when one of Mr. QRS' old frat brothers from the PVC (Premature Ventricular Contraction) house crashes their place for the weekend.
However relationships can also have more serious discord or blocks between them:
- A first degree block places some distance between the two lovers, but still they remain together.
- A second degree type I block is demonstrated by P and QRS Wenkebach who grow steadily further apart until QRS cheats and P dumps the bum, but she forgives him and takes him back only for the cycle to repeat.
- The more dastardly second degree type II occurs in the relationship between P and QRS Mobitz II, where despite the illusion of a stable relationship, QRS gets drunk, cheats, and they break up. P is still a forgiving lady and takes him back, but the scalawag QRS will do it again.
- Divorce, or the third degree block, has complete disassociation between P and QRS
The thin and rhythmless rapper A fib has so many ladies (his P's) that he is only erratically spotted. He is not to be confused with the equally anorexic hip hop artist A flutter who usually dates only twins and triplets. Neither one of these artists jams should be confused with the sudden vibrant beat of the techno SVT, he brings a rapid rhythm that looks almost sinus. Rumor has he too has an eating disorder. Everyone knows a player like their friend Mat (Multifocal Atrial Tachycardia), he's got a different P every week.
Not so for the amphetamine and steroid abusing loner V Tach, no Ps for him in his destructive quest. He used to be a founding member of the all male deathmetal band V Fib, but they were too chaotic even for him.
Monday, June 13, 2011
|Shock ||Hypovolemia +/- electrolyte abnormalities ||Electrolyte abnormalities +/- hypovolemia ||NPO|
|Choice of fluids|
| ||Typically composition is directed by published guidelines (as in DKA) or fluids can be formulated by the considerations for maintenance fluids given in the column furthest to the right. Remember that 40 mEq of potassium in a peripheral line has a maximal infusion rate of 125 mL/hr||Hypovolemic hyponatremia ||Based on electrolytes, specifically: sodium, potassium, chloride, bicarbonate, and glucose (and occasionally phosphate) as well as blood pressure. If they are hypertensive consider hypotonic (1/2 NS) rather than isotonic (NS, LR) solutions even if they are mildly hyponatremic. |
|20 mL/kg of crystalloid (NS or LR) over 15-20 minutes (thus a pressure bag is needed)||100-1000 mL/hr titrated to whatever volume deficit you are correcting||Hypovolemic hyponatremia ||40-20-10 "rule" (for patients with normal electrolyte hemostatic mechanisms) |
|How do you know it's working|
| || ||Hypovolemic hyponatremia ||The patient remains hemodynamically stable without electrolyte abnormalities or worsening renal function|
- Assuming NaHCO3 comes in a stock solution of 1 mEq/mL, then [C] = solute osmolarity / solution volume
= solute osmolarity / (solute volume + solvent volume)
= 150 mEq / (0.150 L + 1 L) = 130 mEq NaHCO3/L
In 850 mL of solvent, then [C] = 150 / 1000 = 150 mEq Na HCO3/L
- Assuming total body water [L] = 0.5 x wgt [kg] (acutally TBW fraction varies between 0.45 and 0.6 depending on gender and age). We want to change the current plasma sodium by 10 mEq/L/24 hours (i.e. the maximum safe change in serum sodium that will not precipitate central pontine myelinolysis) and recall that the concentration of sodium in 0.9% saline or NS is 154 mEq/L, then:
infusion rate = 0.5 x wgt [kg] x 10 / 154 (L x mEq/L/24 hr) / (mEq/L)
= 5 x 1000 x wgt [kg] / (154 x 24) mL/hr
= 1.35 x wgt [kg] mL/hr for NS
If we are using 3% saline, multiply by 0.9/3 (0.9% / 3%) or 0.3
In the case of hypernatremia, we know that 1/2 NS is half the concentration of NS, such that if we did a similar infusion rate calculation the denominator would be half what is above, or simply multiplying by 2.
Sunday, June 12, 2011
|Etiology||Physical Exam||Bedside Diagnostic Tests||History||STAT Diagnostic Tests|
Lines (vascular, GI, GU)
Vomiting / Diarrhea
UA and urine c/s
C. difficile toxin
Side effects of and withdrawal from drugs
|Naloxone trial||MAR (narcotics, corticosteroids, CNS agents)|
Review home medications
Electrolyte abnormalities particularly hypo- and hyperosmolar states
Inadequate pain control
Cardiac arrhythmia and ischemia
Jugular venous pulsation
Heart murmurs and additional sounds
Acute blood loss
|Fecal occult blood test|
(ABG also has an H&H)
Hypoxemia or hypercarbia
Acidosis or alkalosis
|ABG||History of hypercarbia|
Hypo- and hyperglycemia
Insulin or oral antihyperglycemic medications
Intracranial pathology (i.e. stroke, mass effect, traumatic brain injury)
|Neurological examination||Documentation of seizure-like activity without formal history of seizure|
History of dementia or seizure
|CT head without contrast|