Renal impairment affects at least 10 % of community-dwelling and 40 % of long-term care facility adults ≥65 years and older [12, 13]. Kidney function declines with normal senescence—a 30 % decrease in glomerular filtration rate (GFR) occurs between the ages of 30–80 years—as well as other factors such as nephrotoxic medications and comorbidity complications (e.g., diabetes, hypertension) [14, 15]. Since over 25 % of all human medications have been estimated to be mostly excreted by the kidneys, the prescribing practitioner should be aware when renal dose adjustments are necessary (see Table 1) [4].

A major challenge is determining whether older adults have clinically significant renal disease. Absolute serum creatinine level and/or blood urea nitrogen (BUN) have conventionally been used by clinicians as surrogate markers of renal function. These markers, however, are inaccurate in older adults [16]. Very old patients, particularly frail patients, tend to have decreased muscle mass and creatinine production [14]. Thus, some older adults might have a deceptively “normal” serum creatinine when in fact there is renal insufficiency. Serum creatinine and blood urea nitrogen (BUN) in isolation should not be used to estimate renal function (LOE IV) [16].

Creatinine clearance has often been calculated using the Cockroft–Gault equation as an estimate of GFR. This method has traditionally been used by the Food and Drug Administration (FDA) for renal-dose adjustment recommendations [17]. However, this equation’s accuracy, particularly when used for older patients, was called into question [18]. Furthermore, modern creatinine assays are different from when the Cockroft–Gault equation was developed. As a result, this equation might lead to the overestimation of renal function by up to 40 % [17].

There are several currently used or newer measures of renal function. The estimated GFR (eGFR) using the Modification of Diet in Renal Disease (MDRD) formula is reasonably accurate for most non-acute patients, as long as the patient’s weight is not extreme (e.g., amputee, obese, frail) (LOE IIB) [15, 19]. Further evidence is needed before CKD-EPI and cystatin C will be feasible, widely accepted alternatives to MDRD for older patients (LOE III) [15, 19]. A 24-h urine creatinine clearance collection, despite practical limitations in accurate collection, can be considered in some patients (e.g., extremely high or low weight or amputees) who are taking medicines with a narrow therapeutic window (e.g., methotrexate) (LOE IV) [17].

Aging is generally associated with decreased homeostatic functional reserve [9]. The extent of this decrease, which depends on individual genetic makeup, environmental exposures, and comorbidities, can vary greatly between patients. Very old and frail patients are at particularly high risk for experiencing cardiovascular, urinary, and neuropsychiatric side effects [20]. The so-called geriatric syndromes – symptoms that are commonly but inaccurately attributed to the “normal” aging process (e.g., incontinence, falls, or delirium) – are considered by some experts to be preventable patient safety events that are usually caused by adverse medication reactions [21, 22]. However, if a patient has multiple comorbidities that overlap with potential medication side effects, it can be particularly challenging to determine whether a patient’s underlying comorbidity is truly progressing or whether iatrogenesis (medication side effect) has occurred [23]. This section discusses common examples of how geriatric syndromes can be caused by dermatologic medications. Special considerations for anticipating and managing immunosuppressant side effects in older adults are also presented.

Geropharmacology might seem far removed from day-to-day clinical reality, but the aforementioned principles have several important implications for commonly prescribed dermatologic medications. While an exhaustive overview of all dermatologic medications is beyond the scope of this chapter, common examples of pitfalls and a number of prescribing pearls are listed in Tables 1, 2, and 3.

Systemic steroids have a plethora of side effects, including opportunistic infections, ulcers, hyperglycemia, hypertension, delirium, glaucoma, heart failure, weakness from myopathy, arrhythmias, cataracts, osteopenia, and osteoporosis [32]. Although these side effects are not geriatric syndromes per se, the same principles of cautious prescribing and close monitoring also apply to systemic corticosteroids. Careful patient education and coordination of care with the primary care provider or relevant specialists (e.g., cardiologist, ophthalmologist, diabetologist) are recommended prior to and during systemic steroid treatment (LOE IV).

Since hip fracture in older adults is associated with a fivefold increased 3-month mortality risk and morbidity compared to those without fracture, it is important to understand osteoporosis risk and prevention strategies in patients taking corticosteroids [33]. A meta-analysis demonstrated that patients taking a prednisone dose as low as 7.5 mg daily had an increased risk of fracture [34]. The American College of Rheumatology (ACR) published updated guidelines in 2010 (see Appendix for risk tables and algorithm), which provide evidence-based recommendations for monitoring and managing patients who are taking prednisone [35]. They recommend stratifying an individual patient’s risk of fracture based on patient characteristics (e.g., alcohol intake, tobacco use, history of fracture, race, gender, height and weight, baseline hip bone density) and anticipated duration and dose of steroid. Based on this calculated risk, the prescribing provider can determine if lifestyle counseling (including vitamin D and calcium doses), prophylaxis with bisphosphonates, and bone density and height monitoring are appropriate (LOE IV).

Recently, there has been controversy over whether corticosteroid monotherapy is associated with an increased risk of peptic ulcers and GI bleeding. Some experts have argued that GI prophylaxis should not be given for glucocorticoid monotherapy, regardless of dose [32]. A systematic review and meta-analysis demonstrated that peptic ulcer risk is indeed increased (odds ratio 1.43, 95% CI 1.22–1.66) for subjects taking corticosteroids compared to placebo [36]. Subgroup analysis suggested that inpatients, but not ambulatory patients, have a statistically significantly increased risk for ulcers. While this study did not make specific recommendations about GI prophylaxis, it acknowledged that most studies included in the analysis did not specify whether subjects took GI prophylaxis and there was no specific subgroup analysis by age. However, the current authors contend that older patients are generally at higher risk for GI bleeding—probably for multiple reasons, including age-associated changes of GI mucosa, prevalence of H. pylori carriage, and anticoagulant and nonsteroidal anti-inflammatory drug (NSAID) use [37]. Furthermore, older adults who experience GI bleeding have higher mortality rates than do younger adults. Therefore, it is the current authors’ recommendation that GI prophylaxis (H2 receptor antagonist or proton pump inhibitor) should be strongly considered in older adult patients who are taking systemic steroids, particularly if they have a history of GI bleeds, heavy alcohol ingestion, or take anticoagulants, aspirin, or NSAIDs (LOE IV). Aspirin or NSAIDs should be suspended, unless there is a major contraindication to doing so (LOE IV).

Infectious disease prophylaxis is important when prescribing chronic systemic steroids or other immunosuppressants, but preventive measures are particularly important in older adults. During normal aging, there is decreased epidermal barrier function against pathogens and impairment of cell-mediated immunity (immunosenescence), which might predispose patients to many skin conditions and infections [31]. Many older adults also have comorbid cardiopulmonary or renal diseases, which further increase immunosuppression [38–41]. Most of the discussion here centers on systemic steroids, though many principles are generalized to other steroid-sparing agents. The authors focus on the more common and serious viral, fungal, and bacterial infections and discuss strategies to mitigate the risk.

Herpes zoster (shingles) tends to occur with aging due to reactivation of dormant varicella zoster virus that occurs as a consequence of immunosenescence [42]. This condition can cause significant morbidity (e.g., postherpetic neuralgia, herpes zoster ophthalmicus) [43]. The risk of zoster is probably increased in patients taking systemic corticosteroids [44]. Zoster vaccination is recommended as part of health-care maintenance for patients 50 years and older and would ideally occur before the need for systemic immunosuppression. However, one study showed that about one-sixth of eligible patients receive the vaccine [45]. According to the Infectious Disease Society of America (IDSA), this live virus vaccine should not be given to patients taking more than prednisone 20 mg daily, methotrexate 0.4 mg/kg/week, or azathioprine 3 mg/kg/day (LOE IV) [46]. If patients are to be prescribed more intensive immunosuppression, the ACR recommends waiting until at least 4 weeks after zoster immunization, when possible (LOE IV) [42]. Immunocompetent household contacts over 60 years of age may be given the vaccine, but they should avoid close contact with the patient if skin lesions develop (LOE IV) [46].

There is a paucity of data for making influenza vaccination recommendations in patients taking chronic steroids or other immunosuppressants. About half of such studies suggest that patients can mount an immunogenic response [47]. Although the specific risk of influenza from chronic steroid or other immunosuppressant use is poorly characterized, the IDSA recommends that patients receive influenza vaccination (only the inactivated intramuscular type, not live attenuated) on an annual basis (LOE IV) [46]. Patients receiving influenza vaccine within 6 months of rituximab are unlikely to seroconvert and unlikely to be harmed, but the IDSA recommends against administering the vaccine in such cases (LOE IV) [46]. Similarly, pneumococcal vaccination is recommended for chronically immunosuppressed patients, though these patients might have a blunted seroconversion titer (LOE IV) [46].

Pneumocystis jirovecii pneumonia (formerly P. carinii or “PCP”, now called PJP) is an opportunistic fungal infection. There are limited data to support PJP prophylaxis in immunocompromised patients without human immunodeficiency virus (HIV). It should be noted that most published recommendations for prophylaxis were based on uncontrolled studies or expert opinion and did not necessarily include patients prescribed immunosuppression for a primary skin condition, and did not specify whether age per se was an independent risk factor for developing PJP. While PJP is known to have potentially adverse outcomes, especially in HIV-negative patients, experts differ in their recommended indications for which patients on chronic immunosuppression should receive prophylaxis [48]. For instance, some recommend that a 16–20 mg daily prednisone dose should warrant PJP prophylaxis (LOE III, IV) [49, 50]. Others suggest that patients should receive prophylaxis if they take 20 mg prednisone daily for over 1 month and have another immunosuppressing comorbidity or medication (LOE III, IV) [51]. Still others contend that chronic steroid patients with certain CD4 or total lymphocyte counts (LOE III), or interstitial lung disease (LOE IV), or chronic anti-TNF therapy alone or in combination with steroids should receive PJP prophylaxis (LOE IV) [52, 53]. The most convincing systematic review and meta-analysis to date by Green et al. suggests that patients with inflammatory and autoimmune dermatologic conditions who take chronic immunosuppressants probably have a PJP risk that does not outweigh potential side effects from prophylaxis [54]. Based on the available literature, the current authors recommend (LOE IV) considering prophylaxis in patients with chronic immunosuppressing condition (e.g., diabetes, cancer, pulmonary disease), or in those taking two or more immunosuppressants (e.g., chronic prednisone bridging to a steroid-sparing immunosuppressant). The preferred prophylaxis regimen is trimethoprim-sulfamethoxazole (LOE IB), although atovaquone (LOE II), dapsone (LOE II) or pentamidine (LOE IV) may be used [50, 51, 54].

Skin cancer risk has been associated with age, skin type, and certain types of immunosuppressant use [55]. Traditionally, the most commonly associated immunosuppressants with skin cancer risk have been azathioprine and cyclosporine based on solid organ transplant patient data [56]. However, data also suggest that anti-TNF agents and chronic prednisone might increase skin cancer risk, although perhaps not as much as the other immunosuppressants [56, 57]. Extrapolating from the above studies, the current authors recommend annual skin checks in patients receiving chronic thiopurine, cyclosporine, or anti-TNF therapy (LOE IV).

Ethical dilemmas are usually thought of as situations that involve grave life or death scenarios. However, there are ethical considerations when one is prescribing to older adults. The four tenets of medical ethics are autonomy, beneficence, non-maleficence, and justice [58]. Autonomy is defined as the right of rational individuals to make an independent and informed decision about their care. Beneficence is defined as acting in the best interest of the patient. Non-maleficence is not inflicting harm, such as avoiding adverse drug events (ADEs). Justice is defined as the fair treatment of individuals and groups and corresponds to the prescriber being a responsible steward of limited health-care resources. These four tenets are used as a framework to discuss dermatology medication management issues that might be encountered when prescribing to older adults.

To what extent should the elderly and/or their caregivers be allowed to manage their medications, especially when age-related issues such as cognitive impairment may be present? Not all patients are equally autonomous, since this is contingent upon decision-making capacity. Local laws vary as to the definition of “capacity to consent.” [59, 60] Generally, legal definitions include the patient’s “ability to understand the relevant information about proposed [treatments], appreciate their situation, use reason to make a decision, and communicate their choice” [59, 60]. Incapacity is uncommon among healthy, community-dwelling older adults but is probably underrecognized in the general geriatric population [60]. In one pooled analysis study, the diagnosis of incapacity was missed by over half of health-care providers [60]. Possible reasons to explain this underrecognition of incapacity might include health providers giving patients the benefit of the doubt, lacking training to assess capacity, or erroneously interpreting patient acquiescence (i.e., agreeing to treatment does not mean the patient understood the implications of their decision).

The prescriber must carefully balance clinical judgment and intended beneficence with patient autonomy. The potential dilemma of medical paternalism is aptly summarized by Fontanella et al. “The issue that professionals face, especially in the geriatric medical context, is how to ethically determine what represents merely adverse personal choice and what is neglect founded in incapacity?” [61]. The dermatology provider often provides longitudinal care for older adult patients and might gain a sense of familiarity with patient and caregiver goals. However, the provider must exercise caution against complacently assuming authority of a patient’s decisions through unintentional paternalism. A later section discusses the dynamic nature of patient goals and prescribing contexts.

A patient’s social support system might be helpful in achieving treatment adherence and health-care access. For instance, older adults might rely on family or friends to counsel them about health-care decisions, to provide transportation, to overcome language or cultural barriers, or to pay for health-care services and medications [62]. However, dilemmas may arise when the patient’s treatment goals differ from that of the people who provide the social support network. That is, if the patient is overly deferential to their social support network for fear of retribution or offending their family and friends, the patient’s autonomy might be decreased as a result.

Is it ethical to ration limited health-care resources by withholding certain treatments from patients based on age? Another example of medical paternalism is ageism, which is a perceived futility of treatment for adults beyond a certain age. Often, ageism stems from the notion that age is a surrogate of life expectancy, which in turn can be inferred to determine whether limited health-care resources should be reserved instead for patients with longer life expectancy [62, 63]. Providers might have justifiable prescribing standards that differ based on the patient’s context, which might include age. As an extreme example, aggressive cutaneous chemopreventive measures may not be wanted or needed by frail older patients with multiple comorbidities and end-of-life care goals. However, many older adult dermatology patients are often not such outliers, and it is important to avoid ageist prescribing approaches. An individual patient’s life span and quality of life, both of which may be improved by prescriptions, are not necessarily reflected by actuarial statistics that provide information about population-based life expectancy [64]. Thus, the current authors recommend that prescription decisions are made after considering the individual patient’s context (i.e., not only chronologic age) (LOE IV). The authors also recommend negotiating realistic treatment goals with the patient that are within the confines of limited health-care resources. Potential risks and benefits of each intervention should be considered (LOE IV).

As a corollary to this point, older patients, particularly the frail, have been excluded or underrepresented in most clinical trials [65]. Older patients might be at risk for medication interactions or adverse effects due to comorbidities or functional limitations [62, 63]. However, without adequate safety and dosing data, it can be difficult to assess the cost-effectiveness or potential harms in treating older patients [62, 63]. The tenets of justice, beneficence, and non-maleficence are in conflict due to this evidence gap.

In summary, geriatric ethical dilemmas may arise anywhere from pharmaceutical research and development exclusion criteria all the way to the prescriber’s discussion and selection of medications. The above examples highlight the delicacy that is required in balancing autonomy, beneficence, non-maleficence, and justice as well as the complex interplay that exists among the individual patient, caregivers, and society. Later sections further illustrate these ethical issues in geriatric prescribing.

A significant economic burden and potentially negative patient outcomes can result from medication management errors. This topic is particularly relevant to the geriatric population, which consumes about 42 % of total drug expenditures and is at higher risk for developing ADEs [3, 66]. Hospitalization rates, costs, and length of stay are all increased by ADEs [67]. The total cost of preventable ADEs has been estimated at more than $100 billion per year, or around 10 % of annual American health-care expenditures [68]. While there are several published definitions and categories of medication errors, the authors have adapted a framework by Gupta et al. [69]. Medication errors, or unintended consequences of prescription medications, include inappropriate use, overuse, underprescribing, and medication nonadherence. Previously discussed geropharmacological and ethical principles are applied to the description of these four types of medication errors.